WO2024045601A1 - Ironing method and device for garment steamer, and garment steamer, storage medium and electronic device - Google Patents

Ironing method and device for garment steamer, and garment steamer, storage medium and electronic device Download PDF

Info

Publication number
WO2024045601A1
WO2024045601A1 PCT/CN2023/085376 CN2023085376W WO2024045601A1 WO 2024045601 A1 WO2024045601 A1 WO 2024045601A1 CN 2023085376 W CN2023085376 W CN 2023085376W WO 2024045601 A1 WO2024045601 A1 WO 2024045601A1
Authority
WO
WIPO (PCT)
Prior art keywords
ironing
mark
recognition
image
garment
Prior art date
Application number
PCT/CN2023/085376
Other languages
French (fr)
Chinese (zh)
Inventor
潘威滔
Original Assignee
海尔优家智能科技（北京）有限公司
青岛海尔科技有限公司
海尔智家股份有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 海尔优家智能科技（北京）有限公司, 青岛海尔科技有限公司, 海尔智家股份有限公司 filed Critical 海尔优家智能科技（北京）有限公司
Publication of WO2024045601A1 publication Critical patent/WO2024045601A1/en

Links

Classifications

    • D—TEXTILES; PAPER
    • D06—TREATMENT OF TEXTILES OR THE LIKE; LAUNDERING; FLEXIBLE MATERIALS NOT OTHERWISE PROVIDED FOR
    • D06F—LAUNDERING, DRYING, IRONING, PRESSING OR FOLDING TEXTILE ARTICLES
    • D06F73/00—Apparatus for smoothing or removing creases from garments or other textile articles by formers, cores, stretchers, or internal frames, with the application of heat or steam 
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/0464—Convolutional networks [CNN, ConvNet]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
    • G06V30/40—Document-oriented image-based pattern recognition
    • G06V30/41—Analysis of document content
    • G06V30/413—Classification of content, e.g. text, photographs or tables
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
    • G06V30/40—Document-oriented image-based pattern recognition
    • G06V30/42—Document-oriented image-based pattern recognition based on the type of document

Definitions

  • the present application relates to the field of smart home technology, and in particular to a garment ironing method and device, a garment steamer, a storage medium and an electronic device.
  • the object detection method for washing and ironing labels for garment ironing machines mainly involves box and category labeling of the washing and ironing label training data, and then training the model through the object detection method in deep learning.
  • the quantity distribution of each category of washable labels was very different.
  • most categories did not exceed hundreds of annotations.
  • the target detection algorithm finally applied to the garment ironing machine could not accurately identify the ironing machine. Hot information.
  • This application provides an ironing method with a garment steamer, including:
  • ironing information recognition is performed on the washable mark image, and the corresponding one or more ironing parameter recognition methods are obtained based on the ironing information recognition results to pass the required Use the ironing parameter identification method described above to obtain the ironing parameters of the garment ironing machine;
  • an ironing mode corresponding to the garment ironing machine is generated, and the garment ironing machine is controlled according to the ironing mode.
  • This application also provides an ironing device for a garment steamer, including:
  • the clothing maintenance implementation type detection module is used for image recognition of washed mark images and target detection based on the image recognition results;
  • the ironing information detection module is used to identify the ironing information of the washable mark image when it is determined that the clothing maintenance implementation type is an ironable type, and obtain the corresponding one or more ironing information based on the ironing information identification results.
  • An ironing parameter identification method to obtain the ironing parameters of the garment ironing machine through the ironing parameter identification method;
  • a control module configured to generate an ironing mode corresponding to the garment ironing machine based on the ironing parameters, and control the garment ironing machine according to the ironing mode.
  • the present application also provides a computer-readable storage medium.
  • the computer-readable storage medium includes a stored program, wherein when the program is run, it executes and implements any one of the garment ironing methods described above.
  • the present application also provides an electronic device, including a memory and a processor.
  • a computer program is stored in the memory, and the processor is configured to implement any one of the above mentioned garment ironing machines by executing the computer program. method.
  • Figure 1 is a schematic diagram of the hardware environment of an interaction method for smart devices according to an embodiment of the present application
  • FIG. 2 is a schematic flow chart of the garment ironing method provided by this application.
  • FIG. 3 is a schematic structural diagram of the ironing device of the garment steamer provided by this application.
  • Figure 4 is a schematic structural diagram of the electronic device provided by the present application.
  • FIG. 5 is a schematic structural diagram of the garment steamer provided by this application.
  • the object detection method for washing and ironing labels for garment ironing machines mainly involves labeling the washing and ironing label training data with frames and categories, and then training the model through object detection methods in deep learning, such as ResNet, ResNext and ResNest, etc.
  • object detection methods in deep learning such as ResNet, ResNext and ResNest, etc.
  • each water-washed label category is required to have at least thousands of labeled data, and the number between categories must be relatively balanced.
  • the number of high-frequency categories can reach thousands, but the number of labels for most categories does not exceed hundreds.
  • the tail effect is serious, resulting in the target detection algorithm finally applied to the garment ironing machine unable to accurately identify the ironing information.
  • a garment ironing method is provided.
  • This garment ironing method is widely used in whole-house intelligent digital control application scenarios such as Smart Home, smart home, smart home equipment ecology, and smart residence (Intelligence House) ecology.
  • the above-described garment ironing method can be applied to a hardware environment composed of a terminal device 102 and a server 104 as shown in FIG. 1 .
  • the server 104 is connected to the terminal device 102 through the network and can be used to provide services (such as application services, etc.) for the terminal or the client installed on the terminal. It can be on the server or independently.
  • a database is set up on the server to provide data storage services for the server 104.
  • Cloud computing and/or edge computing services can be configured on the server or independently of the server to provide data computing services for the server 104.
  • the above-mentioned network may include but is not limited to at least one of the following: wired network, wireless network.
  • the above-mentioned wired network may include but is not limited to at least one of the following: wide area network, metropolitan area network, and local area network.
  • the above-mentioned wireless network may include at least one of the following: wireless fidelity (Wireless Fidelity, WIFI) and Bluetooth.
  • the terminal device 102 may be, but is not limited to, a PC, a mobile phone, a tablet, a smart air conditioner, a smart hood, a smart refrigerator, a smart oven, a smart stove, a smart washing machine, a smart water heater, a smart washing equipment, a smart dishwasher, or a smart projection device.
  • smart TV smart clothes drying rack, smart curtains, smart audio and video, smart sockets, smart audio, smart speakers, smart fresh air equipment, smart kitchen and bathroom equipment, smart bathroom equipment, smart sweeping robot, smart window cleaning robot, smart mopping robot, Smart air purification equipment, smart steamers, smart microwave ovens, smart kitchen appliances, smart purifiers, smart water dispensers, smart door locks, etc.
  • FIG 2 is a schematic flow chart of the garment ironing method provided by the present application. As shown in Figure 2, the application provides a garment ironing method, which includes the following steps.
  • Step 101 Perform image recognition on the water-washed mark image, and perform target detection based on the image recognition results.
  • the washing mark image of the target clothing can be obtained first through the camera installed on the garment steamer or the user's mobile terminal (for example, mobile phone and tablet computer), and then the collected washing mark image can be initially identified. That is, identification of clothing maintenance methods.
  • the washed mark image can be recognized through a pre-trained target recognition model (i.e., the first recognition model).
  • the model only needs to use two types of ironing methods: ironable and non-ironable.
  • the target is trained to quickly complete the training of the model, so that the trained model can be After obtaining the washable mark image, you only need to output the result of whether it can be ironed or not.
  • Step 102 When it is determined that the clothing maintenance implementation type is an ironable type, ironing information identification is performed on the washable mark image, and one or more corresponding ironing parameter identification methods are obtained based on the ironing information identification results. Through the ironing parameter identification method, the ironing parameters of the garment ironing machine are obtained.
  • step 101 when it is determined that the clothing maintenance implementation type of the target clothing is an ironable type, it is necessary to further identify the ironing text information of the wash mark image. Since the training samples of existing ironing information recognition models are mainly trained on sample images marked with ironing type labels, these models have poor recognition effects on the content of ironing text information. In addition, there are many problems with ironing text information. The number of sample images is small, and the small number used in the general training process results in the model being unable to accurately identify the ironing information in such images. Therefore, in this application, after determining that the clothing maintenance implementation type is an ironable type, it is necessary to identify the ironing information of the washed mark image, that is, to identify the text information in the washed mark image.
  • the recognition result does not contain text Information (that is, it does not contain descriptive information about ironing, such as ironing temperature parameters and ironing duration, etc.).
  • these ironing icons that contain text description content, they can be defined as low-frequency categories.
  • the number of such ironing icons either on the clothes or in the training samples
  • the ironing icon can be recognized through the trained ironing icon recognition model (i.e., the second recognition model), thereby obtaining the ironing icon corresponding to the target clothing.
  • the input data can be considered as high-frequency classification data, that is, this type of data belongs to the model training stage.
  • the ironing information recognition result contains text content
  • the text content and the ironing icon can be recognized separately.
  • Parameter recognition results for example, the ironing icon recognition result is that high temperature ironing is possible, and the text content recognition result is 200 degrees Celsius. Based on these two recognition results, the gear of the garment ironing machine can be adjusted to the high temperature gear first, and then the ironing The temperature is set at 200 degrees Celsius.
  • Step 103 Generate an ironing mode corresponding to the garment ironing machine based on the ironing parameters, and control the garment ironing machine according to the ironing mode.
  • the ironing icon is then used to identify the model.
  • the model recognizes the ironing icon again, thereby outputting the ironing parameters corresponding to the ironing icon, and then sends the ironing parameters to the control unit of the garment ironing machine, thus completing the ironing control of the garment ironing machine.
  • the text information in the ironing icon can be matched with the database through a previously constructed mapping rule database, and the corresponding mapping result can be obtained, so that the ironing
  • the icon recognition model and the mapping rule database will simultaneously send the output results obtained by each to the garment steamer to control the ironing mode of the garment steamer.
  • the output result of the ironing icon recognition model is the maximum ironing temperature. is 110 degrees Celsius (there is a " ⁇ " symbol in the corresponding ironing icon, which is a black solid dot in the ironing icon), and the text content below the ironing icon can be generated through the mapping rule database.
  • Ironing duration parameters (such as text information about the duration below the ironing icon), when the garment ironing machine receives the above two ironing parameters, it will automatically switch to the corresponding ironing mode.
  • the garment ironing method provided by this application first identifies the clothing maintenance method on the wash mark image. When it is determined that the clothing maintenance implementation type is an ironable type, the ironing information is then identified on the wash mark image, and according to the ironing The information recognition results are used to generate ironing parameters for controlling the garment ironing machine. Compared with the existing technology, multi-stage detection and recognition of washed mark images is performed to avoid the problem of poor recognition effect of the target detection model due to too few training samples. Improved the recognition accuracy of ironing information in washed mark images.
  • the method further includes:
  • the clothing maintenance implementation type is a non-ironable type
  • shape recognition is performed on the washable mark in the washable mark image, and based on the shape recognition result, it is determined whether the washable mark in the washable mark image is non-ironable. icon;
  • the washable mark is not a non-ironable icon, determine the clothing maintenance disabling information corresponding to the washable mark based on the shape recognition result of the washable mark, and display the clothing maintenance disabling information through the garment ironing machine, wherein, the clothing maintenance disabling information includes one or more of the following: no washing, no bleaching, and no flipping and drying.
  • the model i.e., the first recognition model
  • the number of samples in this category that is, non-ironable icons
  • the model will Do not dry clean, do not bleach and do not Maintenance implementation types such as reversible drying are incorrectly identified as non-ironable. This is because the above icons contain an " ⁇ " graphic, causing the model to incorrectly identify them.
  • the output result can be defined as a pending result, that is, it is determined to be an unimplementable type, and further judgment is required.
  • the unimplementable information of clothing maintenance is determined based on the shape of the washable mark, such as unwashable or unbleachable, and corresponding information is generated, such as voice information and Text information is displayed through the voice module or display module of the garment steamer.
  • determining whether the washable mark in the washable mark image is a non-ironable icon based on the shape recognition result includes:
  • washable mark image does not include any preset washable mark shape, it is determined that the washable mark in the washable mark image is a non-ironable icon, wherein the preset washable mark shapes include circles, triangles and rectangles. one or a combination of several.
  • the washable mark recognized in the washable mark image is tentatively designated as a non-ironable icon, and it is necessary to perform Hough circle detection on the washable mark, Triangle detection and rectangle detection, when it is determined that there are no circles (corresponding to dry cleaning), triangles (corresponding to bleaching) and rectangles (corresponding to flipping and drying), it can be determined that the unenforceable type of the washable mark is a non-ironable icon, and then the corresponding Notification information such as voice information and text information that cannot be ironed is sent to the garment ironing machine or the user's mobile terminal to inform the user.
  • the wash mark represents other types of clothing care that cannot be performed.
  • the image recognition of water-washed mark images includes:
  • a first recognition model is constructed by training a convolutional neural network.
  • the training process of this model only needs the two ironing icons, which are ironable and non-ironable, as sample data, thereby quickly completing the model training.
  • the first recognition model is trained through the following steps:
  • each sample washing mark image with a corresponding clothing care implementation type label to construct a training sample set, wherein the clothing care implementation type label at least includes an ironable label and a non-ironable label;
  • the convolutional neural network is trained through the training sample set to obtain the first recognition model.
  • the model has higher accuracy in identifying this category; for non-ironable icons, although the number of sample data of this type is small , but this application will later detect the non-ironable icon recognized by the first recognition model through shape detection, thereby improving the accuracy of identifying the ironing mark.
  • ironing information recognition is performed on the washable mark image, and the corresponding one or more items are obtained based on the ironing information recognition result.
  • An ironing parameter identification method to obtain the ironing parameters of the garment ironing machine through the ironing parameter identification method including:
  • the recognition result of the washed mark text content does not contain any text information
  • input the washed mark image into the second recognition model to identify the ironing information, and obtain the ironing icon information output by the second recognition model,
  • the second recognition model is trained by sample washed mark images of ironing icon labels
  • the ironing parameters of the garment ironing machine are generated.
  • optical character recognition Optical Character Recognition, OCR
  • OCR optical Character Recognition
  • the ironable icon is recognized again through the second recognition model to obtain the ironing icon information.
  • the ironing icon includes two " ⁇ " symbols, and the output result of the second recognition model is the highest ironing
  • the temperature is 150 degrees Celsius. Due to the large number of samples of this type of ironing icon, the second recognition model is basically After completing training on these sample data, the recognition task can be completed accurately.
  • the method further includes:
  • the recognition result of the washed mark text content contains text information, perform category mapping on the recognition result of the washed mark text content and the preset ironing rules to obtain the corresponding ironing parameters of the garment steamer.
  • category mapping is performed based on the preset mapping rules of OCR regular expressions.
  • the recognition result of the washed mark text content can be mapped to the corresponding category through the preset ironing rules, the ironing parameters are obtained. If the mapping cannot be performed, It means that the text content is not related to ironing.
  • the method further includes:
  • the ironing parameters of the garment ironing machine are obtained.
  • the washed mark image contains text information
  • two ironing parameter identification methods can be combined, that is, the recognition results of the washed mark text content and the preset ironing rules are mapped to categories, and the washed mark image is The image is input into the second recognition model for ironing information recognition, thereby obtaining the recognition results corresponding to the two ironing parameter recognition methods. Finally, the two recognitions are combined and merged to generate the ironing mode corresponding to the garment ironing machine.
  • the method further includes:
  • An image cropping process is performed on the area range corresponding to the water-washed mark in the water-washed mark image to obtain a target image, so as to perform ironing information recognition or shape recognition on the target image.
  • the area range of the ironing icon in the washed mark image can be obtained based on the results output by the first recognition model, and then the area range is processed Crop, and use the image containing only the ironing icon content as the target image for subsequent target recognition detection to reduce interference during the recognition process.
  • Step 101 Divide all obtained washing and ironing icon training data into two categories: OK Ironing and non-ironable (the difference lies in whether there is a " ⁇ " symbol in the middle of the ironing icon), and then marked according to the target detection method, and trained through the target detection method to obtain two categories for identifying ironable/non-ironable The first recognition model of the target.
  • Step 102 Recognize the washable label image through the first recognition model. If the recognition result is that it cannot be ironed, go to step 103; if the recognition result is that it can be ironed, go to step 106.
  • Step 103 Crop the non-ironable icon into a small image (i.e. target image) according to the frame coordinates (x, y, w, h), and then perform traditional circle detection, triangle detection and rectangle detection on the small image. If there are any of the above If one of the shapes exists, go to step 104; if none of the above shapes exists, go to step 105.
  • Step 104 if the detected thumbnail is an icon that cannot be dry cleaned, cannot be bleached or cannot be flipped and dried, that is, the thumbnail contains a circle, triangle or rectangle, it means that the icon is not a non-ironable icon, so the icon is not output. information, there is no need to adjust the ironing mode of the garment steamer.
  • Step 105 If the detected thumbnail is indeed a non-ironable icon, go to step 111.
  • Step 106 Cut the ironable icon into a small picture according to the frame coordinates (x, y, w, h), and perform OCR recognition on the small picture. Two situations will occur: OCR recognizes text (including digital content), and then go to step 106. 107; OCR cannot recognize any text, go to step 110.
  • Step 107 Perform category mapping according to the mapping rules of the OCR regular expression. There may be two situations: if it can be mapped to a category, go to step 108; if it cannot be mapped to the corresponding category, go to step 109.
  • Step 108 Generate corresponding ironing parameters based on the mapped categories. For example, text content of 80-120 degrees Celsius can be mapped to 80-120 temperature parameters (which are low-frequency categories that can be output). Go to step 111.
  • Step 109 For categories that do not exist in the mapping rules, there is no need to generate output results. For example, if the text content is "199" and cannot be matched to any category through the regular expression, no output will be generated.
  • Step 110 When OCR cannot recognize any text, the ironable icon is directly recognized through the second recognition model and the corresponding ironing parameters are output.
  • Step 111 Combine the output results of step 108 and step 110, and send them to the garment steamer to execute the corresponding ironing mode.
  • the garment ironing device provided by the present application will be described below.
  • the garment ironing device described below and the garment ironing method described above may be mutually referenced.
  • FIG 3 is a schematic structural diagram of an ironing device for a garment steamer provided by this application.
  • this application provides an ironing device for a garment steamer, including a clothing maintenance implementation type detection module 301 and an ironing information detection module. 302 and control module 303, wherein the clothing maintenance implementation type detection module 301 is used to perform image recognition on the washing mark image and perform target detection based on the image recognition results; the ironing information detection module 302 is used to determine that the clothing maintenance implementation type is ironable.
  • the clothing maintenance implementation type detection module 301 is used to perform image recognition on the washing mark image and perform target detection based on the image recognition results;
  • the ironing information detection module 302 is used to determine that the clothing maintenance implementation type is ironable.
  • ironing type perform ironing information identification on the washed mark image, and obtain one or more corresponding ironing parameter identification methods according to the ironing information identification results, so as to obtain the hanger through the ironing parameter identification method.
  • Ironing parameters of the ironing machine the control module 303 is configured to generate
  • the garment ironing device provided by this application first identifies the clothing maintenance method on the wash mark image. When it is determined that the clothing maintenance implementation type is an ironable type, it then identifies the ironing information on the wash mark image and performs the ironing process according to the ironing method. The information recognition results are used to generate ironing parameters for controlling the garment ironing machine. Compared with the existing technology, multi-stage detection and recognition of washed mark images is performed to avoid the problem of poor recognition effect of the target detection model due to too few training samples. Improved the recognition accuracy of ironing information in washed mark images.
  • the device further includes a washing mark shape detection module and a clothing care prompt module, wherein the washing mark shape detection module is used to determine whether the clothing maintenance implementation type is a non-ironable type.
  • the wash mark in the wash mark image is subjected to shape recognition, and based on the shape recognition result, it is determined whether the wash mark in the wash mark image is a non-iron icon; the clothing care prompt module is used to determine if the wash mark is not a non-iron icon.
  • Icon determines the clothing maintenance unimplementable information corresponding to the wash mark, and display the clothing maintenance unimplementable information through the garment steamer, wherein the clothing maintenance unimplementable information Including one or more of do not wash, do not bleach and do not tumble dry.
  • the washable mark shape detection module is specifically used to determine that the washable mark in the washable mark image is non-ironable if the washable mark image does not include any preset washable mark shape. icon, wherein the preset washable mark shape includes one or a combination of circles, triangles and rectangles.
  • the clothing maintenance implementation type detection module includes an image The acquisition unit and the first recognition unit, wherein the image acquisition unit is used to acquire the wash mark image of the target clothing; the first recognition unit is used to input the wash mark image into the first recognition model to identify the clothing maintenance method, and obtain the The clothing care implementation type output by the first recognition model, wherein the first recognition model is trained by sample washing mark images marked with clothing care implementation type labels.
  • the device further includes a sample image acquisition module, a training set construction module and a training module, wherein the sample image acquisition module is used to obtain multiple sample water-washed mark images; the training set construction module is used to provide each The sample washing mark image is marked with a clothing care implementation type label corresponding to a training sample set, wherein the clothing care implementation type label at least includes an ironable label and a non-ironable label; the training module is used to pass the training The sample set is used to train the convolutional neural network to obtain the first recognition model.
  • the ironing information detection module includes a text recognition unit, a second recognition unit and an ironing parameter generation unit, wherein the text recognition unit is used to determine that the clothing maintenance implementation type is ironable. type, perform text information recognition on the water-washed mark image, and obtain the corresponding water-washed mark text content recognition result; the second recognition unit is used to identify the water-washed mark text content if the water-washed mark text content recognition result does not contain any text information.
  • the mark image is input into the second recognition model for ironing information recognition, and the ironing icon information output by the second recognition model is obtained, wherein the second recognition model is trained by sample washed mark images of the ironing icon label. Obtained; the ironing parameter generating unit is used to generate the ironing parameters of the garment ironing machine according to the ironing icon information.
  • the device further includes an ironing parameter mapping module, which is used to, if the recognition result of the text content of the washed mark contains text information, compare the recognition result of the text content of the washed mark with the preset ironing The rules perform category mapping to obtain the ironing parameters of the garment ironing machine.
  • an ironing parameter mapping module which is used to, if the recognition result of the text content of the washed mark contains text information, compare the recognition result of the text content of the washed mark with the preset ironing The rules perform category mapping to obtain the ironing parameters of the garment ironing machine.
  • the ironing information detection module is also used to: when it is determined that the text content recognition result of the water-washed mark contains text information, compare the recognition result of the text content of the water-washed mark with the preset
  • the ironing rules perform category mapping, and the washed mark image is input into the second recognition model to identify the ironing information; based on the category mapping results and the ironing icon information, the ironing parameters of the garment ironing machine are obtained.
  • the device further includes an image acquisition module for The area range corresponding to the water-washed mark in the water-washed mark image is subjected to image cropping processing to obtain a target image, so as to perform ironing information recognition or shape recognition on the target image.
  • FIG. 5 is a schematic structural diagram of the garment steamer provided by this application.
  • this application also provides a garment steamer, including the garment steamer ironing device described in the above embodiments, as well as a steam heater, a body shell, a water tank, a steam nozzle, a steam conduit and a retractable stand.
  • the garment ironing machine uses the ironing device of the garment ironing machine to identify the clothing maintenance method on the washing mark image. When it is determined that the clothing maintenance implementation type is an ironable type, the washing mark image is then identified as ironing information. And based on the ironing information recognition results, ironing parameters for controlling the garment ironing machine are generated.
  • multi-stage detection and recognition of washed mark images is performed to avoid poor recognition results of the target detection model due to too few training samples. It solves the problem of poor quality and improves the recognition accuracy of ironing information in washed mark images.
  • FIG 4 is a schematic structural diagram of an electronic device provided by this application.
  • the electronic device may include: a processor (processor) 410, a communications interface (Communications Interface) 420, a memory (memory) 430 and a communication bus 440.
  • the processor 410, the communication interface 420, and the memory 430 complete communication with each other through the communication bus 440.
  • the processor 410 can call the logical instructions in the memory 430 to execute the garment ironing method.
  • the method includes: performing image recognition on the wash mark image, performing target detection based on the image recognition result; and determining that the clothing maintenance implementation type is ironable.
  • ironing type perform ironing information identification on the washed mark image, and obtain one or more corresponding ironing parameter identification methods according to the ironing information identification results, so as to obtain hanging information through the ironing parameter identification method.
  • Ironing parameters of the ironing machine based on the ironing parameters, generate an ironing mode corresponding to the garment ironing machine, and control the garment ironing machine according to the ironing mode.
  • the above-mentioned logical instructions in the memory 430 can be implemented in the form of software functional units and can be stored in a computer-readable storage medium when sold or used as an independent product.
  • the technical solution of the present application is essentially or the part that contributes to the existing technology or the part of the technical solution can be embodied in the form of a software product.
  • the computer software product is stored in a storage medium, including Several instructions are used to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of this application.
  • the aforementioned storage media include: U disk, mobile hard disk, read-only memory (Read-Only Memory, ROM), random Random Access Memory (RAM), magnetic disks, optical disks and other media that can store program code.
  • this application also provides a computer program product.
  • the computer program product includes a computer program.
  • the computer program can be stored on a computer-readable storage medium.
  • the computer program can execute the above-mentioned steps.
  • the garment ironing method provided by each method includes: performing image recognition on the washable mark image, and performing target detection based on the image recognition result; when it is determined that the clothing maintenance implementation type is an ironable type, performing image recognition on the washable mark image.
  • the image is used to identify ironing information, and according to the ironing information identification result, one or more corresponding ironing parameter identification methods are obtained, so as to obtain the ironing parameters of the garment ironing machine through the ironing parameter identification method; based on the The ironing parameters are used to generate an ironing mode corresponding to the garment ironing machine, and the garment ironing machine is controlled according to the ironing mode.
  • the present application also provides a computer-readable storage medium, which includes a stored program, wherein when the program is run, the garment ironing method provided by each of the above methods is executed, The method includes: performing image recognition on the wash mark image, and performing target detection based on the image recognition result; when determining that the clothing maintenance implementation type is an ironable type, performing ironing information recognition on the wash mark image, and performing ironing information identification on the wash mark image based on the ironing information.
  • Recognition results obtain the corresponding one or more ironing parameter identification methods, so as to obtain the ironing parameters of the garment ironing machine through the ironing parameter identification method; based on the ironing parameters, generate the corresponding information of the garment ironing machine an ironing mode, and control the garment ironing machine according to the ironing mode.
  • the device embodiments described above are only illustrative.
  • the units described as separate components may or may not be physically separated.
  • the components shown as units may or may not be physical units, that is, they may be located in One location, or it can be distributed across multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. Persons of ordinary skill in the art can understand and implement the method without any creative effort.
  • each embodiment can be implemented by software plus a necessary general hardware platform, and of course, it can also be implemented by hardware.
  • the computer software product can be stored in a computer-readable storage medium, such as ROM/RAM, magnetic Discs, compact discs, etc., including Several instructions are used to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in various embodiments or certain parts of the embodiments.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Artificial Intelligence (AREA)
  • General Physics & Mathematics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Computation (AREA)
  • Computing Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Software Systems (AREA)
  • General Health & Medical Sciences (AREA)
  • Multimedia (AREA)
  • Biomedical Technology (AREA)
  • Molecular Biology (AREA)
  • General Engineering & Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Mathematical Physics (AREA)
  • Computational Linguistics (AREA)
  • Biophysics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Databases & Information Systems (AREA)
  • Medical Informatics (AREA)
  • Textile Engineering (AREA)
  • Image Analysis (AREA)
  • Professional, Industrial, Or Sporting Protective Garments (AREA)

Abstract

The present application relates to the technical field of smart homes/intelligent households. Disclosed are an ironing method and device for a garment steamer, and a garment steamer, a storage medium and an electronic device. The ironing method for a garment steamer comprises: performing image recognition on a care label image, and performing target detection on the basis of an image recognition result; when it is determined that a clothes maintenance implementation type is an ironable type, performing ironing information recognition on the care label image, and acquiring one or more corresponding ironing parameter recognition modes according to an ironing information recognition result, so as to obtain ironing parameters of a garment steamer by means of the ironing parameter recognition modes; and on the basis of the ironing parameters, generating an ironing mode corresponding to the garment steamer, and controlling the garment steamer according to the ironing mode.

Description

挂烫机熨烫方法、装置、挂烫机、存储介质及电子装置Garment ironing method, device, garment ironing machine, storage medium and electronic device
相关申请的交叉引用Cross-references to related applications
本申请要求于2022年08月30日提交的申请号为2022110491589，名称为“挂烫机熨烫方法、装置、挂烫机、存储介质及电子装置”的中国专利申请的优先权，其通过引用方式全部并入本文。This application claims priority to the Chinese patent application with application number 2022110491589 and titled "Garment Steamer Ironing Method, Device, Garment Steamer, Storage Medium and Electronic Device" filed on August 30, 2022, which is incorporated by reference All methods are incorporated into this article.
技术领域Technical field
本申请涉及智能家居技术领域，尤其涉及一种挂烫机熨烫方法、装置、挂烫机、存储介质及电子装置。The present application relates to the field of smart home technology, and in particular to a garment ironing method and device, a garment steamer, a storage medium and an electronic device.
背景技术Background technique
针对挂烫机的水洗熨烫标目标检测方法，主要是对水洗熨烫标训练数据进行框和类别标注，然后通过深度学习中的目标检测方法对模型训练。但在实际进行训练数据采集过程中水洗唛各类别之间的数量分布差异极大，同时大部分类别标注未超过上百的标注量，导致最终应用于挂烫机的目标检测算法无法准确识别熨烫信息。The object detection method for washing and ironing labels for garment ironing machines mainly involves box and category labeling of the washing and ironing label training data, and then training the model through the object detection method in deep learning. However, during the actual training data collection process, the quantity distribution of each category of washable labels was very different. At the same time, most categories did not exceed hundreds of annotations. As a result, the target detection algorithm finally applied to the garment ironing machine could not accurately identify the ironing machine. Hot information.
因此，现在亟需一种挂烫机熨烫方法、装置、挂烫机、存储介质及电子装置来解决上述问题。Therefore, there is an urgent need for a garment ironing method, device, garment ironing machine, storage medium and electronic device to solve the above problems.
发明内容Contents of the invention
本申请提供一种挂烫机熨烫方法，包括：This application provides an ironing method with a garment steamer, including:
对水洗唛图像进行图像识别，基于图像识别结果进行目标检测；Perform image recognition on washed mark images and perform target detection based on the image recognition results;
在确定衣物保养实施类型为可熨烫类型时，对所述水洗唛图像进行熨烫信息识别，并根据熨烫信息识别结果，获取对应的一种或多种熨烫参数识别方式，以通过所述熨烫参数识别方式，得到挂烫机的熨烫参数；When it is determined that the clothing maintenance implementation type is an ironable type, ironing information recognition is performed on the washable mark image, and the corresponding one or more ironing parameter recognition methods are obtained based on the ironing information recognition results to pass the required Use the ironing parameter identification method described above to obtain the ironing parameters of the garment ironing machine;
基于所述熨烫参数，生成所述挂烫机对应的熨烫模式，并根据所述熨烫模式控制所述挂烫机。Based on the ironing parameters, an ironing mode corresponding to the garment ironing machine is generated, and the garment ironing machine is controlled according to the ironing mode.
本申请还提供一种挂烫机熨烫装置，包括：This application also provides an ironing device for a garment steamer, including:
衣物保养实施类型检测模块，用于对水洗唛图像进行图像识别，基于图像识别结果进行目标检测； The clothing maintenance implementation type detection module is used for image recognition of washed mark images and target detection based on the image recognition results;
熨烫信息检测模块，用于在确定衣物保养实施类型为可熨烫类型时，对所述水洗唛图像进行熨烫信息识别，并根据熨烫信息识别结果，获取对应的一种或多种熨烫参数识别方式，以通过所述熨烫参数识别方式，获取挂烫机的熨烫参数；The ironing information detection module is used to identify the ironing information of the washable mark image when it is determined that the clothing maintenance implementation type is an ironable type, and obtain the corresponding one or more ironing information based on the ironing information identification results. An ironing parameter identification method to obtain the ironing parameters of the garment ironing machine through the ironing parameter identification method;
控制模块，用于基于所述熨烫参数，生成所述挂烫机对应的熨烫模式，并根据所述熨烫模式控制所述挂烫机。A control module, configured to generate an ironing mode corresponding to the garment ironing machine based on the ironing parameters, and control the garment ironing machine according to the ironing mode.
本申请还提供一种计算机可读的存储介质，所述计算机可读的存储介质包括存储的程序，其中，所述程序运行时执行实现如上述任一种所述挂烫机熨烫方法。The present application also provides a computer-readable storage medium. The computer-readable storage medium includes a stored program, wherein when the program is run, it executes and implements any one of the garment ironing methods described above.
本申请还提供一种电子装置，包括存储器和处理器，所述存储器中存储有计算机程序，所述处理器被设置为通过所述计算机程序执行实现如上述任一种所述挂烫机熨烫方法。The present application also provides an electronic device, including a memory and a processor. A computer program is stored in the memory, and the processor is configured to implement any one of the above mentioned garment ironing machines by executing the computer program. method.
附图说明Description of drawings
此处的附图被并入说明书中并构成本说明书的一部分，示出了符合本申请的实施例，并与说明书一起用于解释本申请的原理。The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the application and together with the description, serve to explain the principles of the application.
为了更清楚地说明本申请实施例或现有技术中的技术方案，下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍，显而易见地，对于本领域普通技术人员而言，在不付出创造性劳动性的前提下，还可以根据这些附图获得其他的附图。In order to more clearly explain the embodiments of the present application or the technical solutions in the prior art, the following will briefly introduce the drawings needed to describe the embodiments or the prior art. Obviously, for those of ordinary skill in the art, It is said that other drawings can be obtained based on these drawings without exerting creative labor.
图1是根据本申请实施例的一种智能设备的交互方法的硬件环境示意图；Figure 1 is a schematic diagram of the hardware environment of an interaction method for smart devices according to an embodiment of the present application;
图2为本申请提供的挂烫机熨烫方法的流程示意图；Figure 2 is a schematic flow chart of the garment ironing method provided by this application;
图3为本申请提供的挂烫机熨烫装置的结构示意图；Figure 3 is a schematic structural diagram of the ironing device of the garment steamer provided by this application;
图4为本申请提供的电子装置的结构示意图；Figure 4 is a schematic structural diagram of the electronic device provided by the present application;
图5为本申请提供的挂烫机的结构示意图。Figure 5 is a schematic structural diagram of the garment steamer provided by this application.
具体实施方式Detailed ways
为了使本技术领域的人员更好地理解本申请方案，下面将结合本申请实施例中的附图，对本申请实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例仅仅是本申请一部分的实施例，而不是全部 的实施例。基于本申请中的实施例，本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例，都应当属于本申请保护的范围。In order to enable those in the technical field to better understand the solutions of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only Examples that are part of this application, not all embodiment. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative efforts should fall within the scope of protection of this application.
需要说明的是，本申请中的术语“第一”、“第二”等是用于区别类似的对象，而不必用于描述特定的顺序或先后次序。应该理解这样使用的数据在适当情况下可以互换，以便这里描述的本申请的实施例能够以除了在这里图示或描述的那些以外的顺序实施。此外，术语“包括”和“具有”以及他们的任何变形，意图在于覆盖不排他的包含，例如，包含了一系列步骤或单元的过程、方法、***、产品或设备不必限于清楚地列出的那些步骤或单元，而是可包括没有清楚地列出的或对于这些过程、方法、产品或设备固有的其它步骤或单元。It should be noted that the terms "first", "second", etc. in this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It is to be understood that the data so used are interchangeable under appropriate circumstances so that the embodiments of the application described herein can be practiced in sequences other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, e.g., a process, method, system, product, or apparatus that encompasses a series of steps or units and need not be limited to those explicitly listed. Those steps or elements may instead include other steps or elements not expressly listed or inherent to the process, method, product or apparatus.
随着智能家居的发展，应用于挂烫机的目标检测技术也越来越多。针对挂烫机的水洗熨烫标目标检测方法，主要是对水洗熨烫标训练数据进行框和类别标注，然后通过深度学习中的目标检测方法对模型训练，例如，ResNet、ResNext和ResNest等，从而能够检测图片中的水洗唛框位置及类别信息，进而再根据类型信息，确定挂烫机当前的熨烫参数。With the development of smart homes, more and more target detection technologies are applied to garment steamers. The object detection method for washing and ironing labels for garment ironing machines mainly involves labeling the washing and ironing label training data with frames and categories, and then training the model through object detection methods in deep learning, such as ResNet, ResNext and ResNest, etc. Thus, the position and category information of the washable mark frame in the picture can be detected, and then the current ironing parameters of the garment steamer can be determined based on the type information.
由于深度学习本身的算法性质，要求每个水洗唛类别至少要达到上千量级的标注数据，而且类别间数量要较为均衡。但在实际进行训练数据采集过程中，经统计发现，水洗唛各类别之间的数量分布差异极大，高频类别可达上千量，但是大部分类别标注未超过上百的标注量，长尾效应严重，导致最终应用于挂烫机的目标检测算法无法准确识别熨烫信息。Due to the algorithmic nature of deep learning itself, each water-washed label category is required to have at least thousands of labeled data, and the number between categories must be relatively balanced. However, during the actual training data collection process, statistics found that the quantity distribution of each category of washed labels is very different. The number of high-frequency categories can reach thousands, but the number of labels for most categories does not exceed hundreds. The tail effect is serious, resulting in the target detection algorithm finally applied to the garment ironing machine unable to accurately identify the ironing information.
因此，现在亟需一种挂烫机熨烫方法、装置、挂烫机、存储介质及电子装置来解决上述问题。Therefore, there is an urgent need for a garment ironing method, device, garment ironing machine, storage medium and electronic device to solve the above problems.
根据本申请实施例的一个方面，提供了一种挂烫机熨烫方法。该挂烫机熨烫方法广泛应用于智慧家庭(Smart Home)、智能家居、智能家用设备生态、智慧住宅(Intelligence House)生态等全屋智能数字化控制应用场景。可选地，在本实施例中，上述挂烫机熨烫方法可以应用于如图1所示的由终端设备102和服务器104所构成的硬件环境中。如图1所示，服务器104通过网络与终端设备102进行连接，可用于为终端或终端上安装的客户端提供服务(如应用服务等)，可在服务器上或独立 于服务器设置数据库，用于为服务器104提供数据存储服务，可在服务器上或独立于服务器配置云计算和/或边缘计算服务，用于为服务器104提供数据运算服务。According to one aspect of the embodiments of the present application, a garment ironing method is provided. This garment ironing method is widely used in whole-house intelligent digital control application scenarios such as Smart Home, smart home, smart home equipment ecology, and smart residence (Intelligence House) ecology. Optionally, in this embodiment, the above-described garment ironing method can be applied to a hardware environment composed of a terminal device 102 and a server 104 as shown in FIG. 1 . As shown in Figure 1, the server 104 is connected to the terminal device 102 through the network and can be used to provide services (such as application services, etc.) for the terminal or the client installed on the terminal. It can be on the server or independently. A database is set up on the server to provide data storage services for the server 104. Cloud computing and/or edge computing services can be configured on the server or independently of the server to provide data computing services for the server 104.
上述网络可以包括但不限于以下至少之一：有线网络，无线网络。上述有线网络可以包括但不限于以下至少之一：广域网，城域网，局域网，上述无线网络可以包括但不限于以下至少之一：无线保真(Wireless Fidelity，WIFI)，蓝牙。终端设备102可以并不限定于为PC、手机、平板电脑、智能空调、智能烟机、智能冰箱、智能烤箱、智能炉灶、智能洗衣机、智能热水器、智能洗涤设备、智能洗碗机、智能投影设备、智能电视、智能晾衣架、智能窗帘、智能影音、智能插座、智能音响、智能音箱、智能新风设备、智能厨卫设备、智能卫浴设备、智能扫地机器人、智能擦窗机器人、智能拖地机器人、智能空气净化设备、智能蒸箱、智能微波炉、智能厨宝、智能净化器、智能饮水机、智能门锁等。The above-mentioned network may include but is not limited to at least one of the following: wired network, wireless network. The above-mentioned wired network may include but is not limited to at least one of the following: wide area network, metropolitan area network, and local area network. The above-mentioned wireless network may include at least one of the following: wireless fidelity (Wireless Fidelity, WIFI) and Bluetooth. The terminal device 102 may be, but is not limited to, a PC, a mobile phone, a tablet, a smart air conditioner, a smart hood, a smart refrigerator, a smart oven, a smart stove, a smart washing machine, a smart water heater, a smart washing equipment, a smart dishwasher, or a smart projection device. , smart TV, smart clothes drying rack, smart curtains, smart audio and video, smart sockets, smart audio, smart speakers, smart fresh air equipment, smart kitchen and bathroom equipment, smart bathroom equipment, smart sweeping robot, smart window cleaning robot, smart mopping robot, Smart air purification equipment, smart steamers, smart microwave ovens, smart kitchen appliances, smart purifiers, smart water dispensers, smart door locks, etc.
在现有的水洗唛熨烫信息识别模型中，采集的水洗熨烫标训练数据类别之间数量差异极大，导致识别模型对于低频数量的类别识别效果极差；并且，在一些识别模型的训练过程中，由于没有对除熨烫标以外的其他类别水洗唛标注，导致和不可熨烫标类似的，包括不可干洗，不可漂白和不可翻转干燥等，当识别模型在对其进行识别时，极易被误识别成不可熨烫标，即模型容易把相似的不可实施类别，误识别为不可熨烫类别。In the existing washed and ironed label information recognition model, the number of collected washed and ironed label training data categories is greatly different, resulting in a very poor recognition effect of the recognition model on low-frequency categories; and, in the training of some recognition models During the process, there was no labeling of other categories of washable labels except ironing labels, which resulted in similar labels to non-ironable labels, including do not dry clean, do not bleach, do not turn over and dry, etc. When the recognition model was identifying them, it was extremely difficult to identify them. It is easy to be misidentified as a non-ironable label, that is, the model easily misidentifies similar non-implementable categories as non-ironable categories.
图2为本申请提供的挂烫机熨烫方法的流程示意图，如图2所示，本申请提供了一种挂烫机熨烫方法，包括下述步骤。Figure 2 is a schematic flow chart of the garment ironing method provided by the present application. As shown in Figure 2, the application provides a garment ironing method, which includes the following steps.
步骤101，对水洗唛图像进行图像识别，基于图像识别结果进行目标检测。Step 101: Perform image recognition on the water-washed mark image, and perform target detection based on the image recognition results.
在本申请中，首先可通过设置在挂烫机上的摄像头，或用户的移动终端(例如，手机和平板电脑)，获取目标衣物的水洗唛图像，进而对采集到的水洗唛图像进行初步识别，即进行衣物保养方式识别。在一实施例中，可通过前期训练好的目标识别模型(即第一识别模型)对水洗唛图像进行识别，该模型在训练过程中，只需利用可熨烫和不可熨烫两种熨烫标进行训练，从而快速完成对模型的训练，使得训练好的模型在 获取到水洗唛图像后，只需输出是否可熨烫或不可熨烫的结果。In this application, the washing mark image of the target clothing can be obtained first through the camera installed on the garment steamer or the user's mobile terminal (for example, mobile phone and tablet computer), and then the collected washing mark image can be initially identified. That is, identification of clothing maintenance methods. In one embodiment, the washed mark image can be recognized through a pre-trained target recognition model (i.e., the first recognition model). During the training process, the model only needs to use two types of ironing methods: ironable and non-ironable. The target is trained to quickly complete the training of the model, so that the trained model can be After obtaining the washable mark image, you only need to output the result of whether it can be ironed or not.
步骤102，在确定衣物保养实施类型为可熨烫类型时，对所述水洗唛图像进行熨烫信息识别，并根据熨烫信息识别结果，获取对应的一种或多种熨烫参数识别方式，以通过所述熨烫参数识别方式，得到挂烫机的熨烫参数。Step 102: When it is determined that the clothing maintenance implementation type is an ironable type, ironing information identification is performed on the washable mark image, and one or more corresponding ironing parameter identification methods are obtained based on the ironing information identification results. Through the ironing parameter identification method, the ironing parameters of the garment ironing machine are obtained.
在本申请中，通过步骤101，在确定目标衣物的衣物保养实施类型为可熨烫类型时，需要进一步对该水洗唛图像进行熨烫文字信息识别。由于现有的熨烫信息识别模型的训练样本主要是标记有熨烫类型标签的样本图像进行训练，导致这些模型对于熨烫文字信息的内容识别效果较差，再加上关于熨烫文字信息的样本图像的数量较少，一般的训练过程中所使用的数量较少，导致模型在对此类图像进行识别时，无法准确识别得到其中的熨烫信息。因此，在本申请中，在确定衣物保养实施类型为可熨烫类型后，需要对水洗唛图像进行熨烫信息识别，即对水洗唛图像中的文字信息进行识别，若识别结果中不包含文字信息(即不包含熨烫方面的描述信息，例如，熨烫温度参数和熨烫时长等，针对这些包含有文字描述内容的熨烫图标，可定义为低频类别，一般这种熨烫图标的数量无论在衣物上或训练时的样本中都较少)，则可通过已训练好的熨烫图标识别模型(即第二识别模型)对熨烫图标进行识别，由此得到目标衣物对应的熨烫参数，由于熨烫图标中不包含关于熨烫参数的文字内容，则通过熨烫图标识别模型进行识别时，输入的数据可认为是高频分类的数据，即此类数据在模型训练阶段是属于数量较多的样本数据，使得模型在对此类图像进行识别时，输出的识别结果可保证较高的准确性。在一实施例中，若熨烫信息识别结果中包含有文字内容，可分别对文字内容和熨烫图标进行识别，通过将两种不同的熨烫参数识别方式组合，得到更为准确的熨烫参数识别结果，例如，熨烫图标识别结果为可高温熨烫，文字内容识别结果为200摄氏度，根据这两个识别结果，可先将挂烫机的档位调整到高温档，再将熨烫温度设置在200摄氏度。In this application, through step 101, when it is determined that the clothing maintenance implementation type of the target clothing is an ironable type, it is necessary to further identify the ironing text information of the wash mark image. Since the training samples of existing ironing information recognition models are mainly trained on sample images marked with ironing type labels, these models have poor recognition effects on the content of ironing text information. In addition, there are many problems with ironing text information. The number of sample images is small, and the small number used in the general training process results in the model being unable to accurately identify the ironing information in such images. Therefore, in this application, after determining that the clothing maintenance implementation type is an ironable type, it is necessary to identify the ironing information of the washed mark image, that is, to identify the text information in the washed mark image. If the recognition result does not contain text Information (that is, it does not contain descriptive information about ironing, such as ironing temperature parameters and ironing duration, etc.). For these ironing icons that contain text description content, they can be defined as low-frequency categories. Generally, the number of such ironing icons (either on the clothes or in the training samples), the ironing icon can be recognized through the trained ironing icon recognition model (i.e., the second recognition model), thereby obtaining the ironing icon corresponding to the target clothing. Parameters, since the ironing icon does not contain text content about the ironing parameters, when identifying through the ironing icon recognition model, the input data can be considered as high-frequency classification data, that is, this type of data belongs to the model training stage. A large amount of sample data enables the model to output recognition results with higher accuracy when identifying such images. In one embodiment, if the ironing information recognition result contains text content, the text content and the ironing icon can be recognized separately. By combining two different ironing parameter recognition methods, a more accurate ironing can be obtained. Parameter recognition results, for example, the ironing icon recognition result is that high temperature ironing is possible, and the text content recognition result is 200 degrees Celsius. Based on these two recognition results, the gear of the garment ironing machine can be adjusted to the high temperature gear first, and then the ironing The temperature is set at 200 degrees Celsius.
步骤103，基于所述熨烫参数，生成所述挂烫机对应的熨烫模式，并根据所述熨烫模式控制所述挂烫机。Step 103: Generate an ironing mode corresponding to the garment ironing machine based on the ironing parameters, and control the garment ironing machine according to the ironing mode.
在本申请中，在确定目标衣物可熨烫后，进而通过熨烫图标识别模 型再次对熨烫图标进行识别，从而输出该熨烫图标对应的熨烫参数，再将熨烫参数发送到挂烫机的控制单元，已完成对挂烫机的熨烫控制。在一实施例中，对于存在熨烫文字信息的熨烫图标，可通过前期构建好的映射规则数据库，将熨烫图标中的文字信息与该数据库进行匹配，获取到对应映射结果，使得熨烫图标识别模型和映射规则数据库将各自获取到的输出结果，同时发送到挂烫机，以用于对挂烫机的熨烫模式进行控制，例如，熨烫图标识别模型输出结果为最高熨烫温度为110摄氏度(对应的熨烫图标中有一个“■”的符号，在熨烫图标中为一个黑色实心点)，而熨烫图标中下方的文字内容，则可通过映射规则数据库生成对应的熨烫时长参数(如该熨烫图标下方有关于时长的文字信息)，当挂烫机接收到上述两种熨烫参数后，自动切换成相应的熨烫模式。In this application, after it is determined that the target clothing can be ironed, the ironing icon is then used to identify the model. The model recognizes the ironing icon again, thereby outputting the ironing parameters corresponding to the ironing icon, and then sends the ironing parameters to the control unit of the garment ironing machine, thus completing the ironing control of the garment ironing machine. In one embodiment, for an ironing icon with ironing text information, the text information in the ironing icon can be matched with the database through a previously constructed mapping rule database, and the corresponding mapping result can be obtained, so that the ironing The icon recognition model and the mapping rule database will simultaneously send the output results obtained by each to the garment steamer to control the ironing mode of the garment steamer. For example, the output result of the ironing icon recognition model is the maximum ironing temperature. is 110 degrees Celsius (there is a "■" symbol in the corresponding ironing icon, which is a black solid dot in the ironing icon), and the text content below the ironing icon can be generated through the mapping rule database. Ironing duration parameters (such as text information about the duration below the ironing icon), when the garment ironing machine receives the above two ironing parameters, it will automatically switch to the corresponding ironing mode.
本申请提供的挂烫机熨烫方法，首先对水洗唛图像进行衣物保养方式识别，在确定衣物保养实施类型为可熨烫类型时，再对水洗唛图像进行熨烫信息识别，并根据熨烫信息识别结果，生成用于控制挂烫机的熨烫参数，相比现有技术，通过对水洗唛图像进行多阶段检测识别，避免目标检测模型由于训练样本过少导致识别效果较差的问题，提高了水洗唛图像中熨烫信息的识别准确率。The garment ironing method provided by this application first identifies the clothing maintenance method on the wash mark image. When it is determined that the clothing maintenance implementation type is an ironable type, the ironing information is then identified on the wash mark image, and according to the ironing The information recognition results are used to generate ironing parameters for controlling the garment ironing machine. Compared with the existing technology, multi-stage detection and recognition of washed mark images is performed to avoid the problem of poor recognition effect of the target detection model due to too few training samples. Improved the recognition accuracy of ironing information in washed mark images.
在上述实施例的基础上，所述方法还包括：Based on the above embodiments, the method further includes:
在确定所述衣物保养实施类型为不可熨烫类型时，对所述水洗唛图像中的水洗唛进行形状识别，并根据形状识别结果，确定所述水洗唛图像中的水洗唛是否为不可熨烫图标；When it is determined that the clothing maintenance implementation type is a non-ironable type, shape recognition is performed on the washable mark in the washable mark image, and based on the shape recognition result, it is determined whether the washable mark in the washable mark image is non-ironable. icon;
若所述水洗唛不是不可熨烫图标，根据所述水洗唛的形状识别结果，确定所述水洗唛对应的衣物保养不可实施信息，并通过所述挂烫机展示所述衣物保养不可实施信息，其中，所述衣物保养不可实施信息包括不可洗涤、不可漂白和不可翻转干燥中的一种或多种。If the washable mark is not a non-ironable icon, determine the clothing maintenance disabling information corresponding to the washable mark based on the shape recognition result of the washable mark, and display the clothing maintenance disabling information through the garment ironing machine, Wherein, the clothing maintenance disabling information includes one or more of the following: no washing, no bleaching, and no flipping and drying.
在本申请中，由于大部分衣物的水洗唛中均为可熨烫图标，一般都是一些材质较为特殊的衣物存在不可熨烫的需求，因此，模型(即第一识别模型)在训练过程中，该类别(即不可熨烫图标)的样本数量存在偏少的情况，导致模型在对这些不可熨烫图标进行识别时，准确率得不到保证，例如，由于样本数量过少，模型会将不可干洗、不可漂白和不 可翻转干燥等类别的保养实施类型，错误的识别为不可熨烫，这是因为上述这些图标中均包含有一个“×”的图形，导致模型会错误识别。因此，当模型输出结果为不可熨烫的衣物保养实施类型时，此时可将该输出结果定义为待定结果，即判定为不可实施类型，需要对其进一步进行判断。进一步地，在判定该水洗唛的不可实施类型不是不可熨烫图标时，进而根据水洗唛形状，确定衣物保养不可实施信息，如不可洗涤或不可漂白等，并生成对应的信息，如语音信息和文字信息，通过挂烫机的语音模块或显示模块进行展示。In this application, since most of the laundry labels of clothes have ironable icons, generally some clothes with special materials have non-ironable requirements. Therefore, during the training process, the model (i.e., the first recognition model) , the number of samples in this category (that is, non-ironable icons) is too small, resulting in the accuracy of the model not being guaranteed when identifying these non-ironable icons. For example, due to too few samples, the model will Do not dry clean, do not bleach and do not Maintenance implementation types such as reversible drying are incorrectly identified as non-ironable. This is because the above icons contain an "×" graphic, causing the model to incorrectly identify them. Therefore, when the model output result is a clothing maintenance implementation type that cannot be ironed, the output result can be defined as a pending result, that is, it is determined to be an unimplementable type, and further judgment is required. Further, when it is determined that the unenforceable type of the washable mark is not the unironable icon, the unimplementable information of clothing maintenance is determined based on the shape of the washable mark, such as unwashable or unbleachable, and corresponding information is generated, such as voice information and Text information is displayed through the voice module or display module of the garment steamer.
在上述实施例的基础上，所述根据形状识别结果，确定所述水洗唛图像中的水洗唛是否为不可熨烫图标，包括：Based on the above embodiment, determining whether the washable mark in the washable mark image is a non-ironable icon based on the shape recognition result includes:
若所述水洗唛图像中不包括任何预设水洗唛形状时，确定所述水洗唛图像中的水洗唛为不可熨烫图标，其中，所述预设水洗唛形状包括圆形、三角形和矩形中的一种或几种的组合。If the washable mark image does not include any preset washable mark shape, it is determined that the washable mark in the washable mark image is a non-ironable icon, wherein the preset washable mark shapes include circles, triangles and rectangles. one or a combination of several.
在本申请中，当第一识别模型输出结果为不可实施类型时，此时，将水洗唛图像中识别的水洗唛暂定为不可熨烫图标，需要通过对该水洗唛进行霍夫圆检测、三角形检测和矩形检测，在确定不存在圆形(对应干洗)、三角形(对应漂白)和矩形(对应翻转干燥)时，可判定该水洗唛的不可实施类型为不可熨烫图标，进而生成对应的不可熨烫告知信息，如语音信息和文字信息，发送到挂烫机或用户移动终端，从而告知给用户。当水洗唛中存在圆形、三角形和矩形中任意一种时，则表示该水洗唛代表其他的衣物保养不可实施类型。In this application, when the output result of the first recognition model is an unimplementable type, at this time, the washable mark recognized in the washable mark image is tentatively designated as a non-ironable icon, and it is necessary to perform Hough circle detection on the washable mark, Triangle detection and rectangle detection, when it is determined that there are no circles (corresponding to dry cleaning), triangles (corresponding to bleaching) and rectangles (corresponding to flipping and drying), it can be determined that the unenforceable type of the washable mark is a non-ironable icon, and then the corresponding Notification information such as voice information and text information that cannot be ironed is sent to the garment ironing machine or the user's mobile terminal to inform the user. When there is any type of circle, triangle or rectangle in the wash mark, it means that the wash mark represents other types of clothing care that cannot be performed.
在上述实施例的基础上，所述对水洗唛图像进行图像识别，包括：Based on the above embodiments, the image recognition of water-washed mark images includes:
获取目标衣物的水洗唛图像；Obtain the washing mark image of the target clothing;
将所述水洗唛图像输入到第一识别模型中进行衣物保养方式识别，得到由所述第一识别模型输出的衣物保养实施类型，其中，所述第一识别模型是由标记有衣物保养实施类型标签的样本水洗唛图像训练得到的。Input the washed mark image into the first recognition model to identify the clothing care mode, and obtain the clothing care implementation type output by the first recognition model, wherein the first recognition model is a clothing care implementation type marked with The labeled sample water-washed label images are trained.
在本申请中，通过对卷积神经网络进行训练，构建第一识别模型，该模型的训练过程只需将可熨烫和不可熨烫这两种熨烫图标作为样本数据，从而快速完成对模型的训练。 In this application, a first recognition model is constructed by training a convolutional neural network. The training process of this model only needs the two ironing icons, which are ironable and non-ironable, as sample data, thereby quickly completing the model training.
在上述实施例的基础上，所述第一识别模型通过以下步骤训练得到：Based on the above embodiments, the first recognition model is trained through the following steps:
获取多个样本水洗唛图像；Obtain water-washed mark images of multiple samples;
为各个所述样本水洗唛图像标记对应的衣物保养实施类型标签，构建得到训练样本集，其中，所述衣物保养实施类型标签至少包括可熨烫标签和不可熨烫标签；Label each sample washing mark image with a corresponding clothing care implementation type label to construct a training sample set, wherein the clothing care implementation type label at least includes an ironable label and a non-ironable label;
通过所述训练样本集，对卷积神经网络进行训练，得到所述第一识别模型。The convolutional neural network is trained through the training sample set to obtain the first recognition model.
在本申请中，由于可熨烫图标的数量较多，使得该模型在对此类别进行识别时，具有较高的准确性；而对于不可熨烫图标，虽然这类型的样本数据的数量较少，但是本申请后续还会通过形状检测，对第一识别模型识别的不可熨烫图标再一次进行检测，从而提高熨烫标的识别准确率。In this application, due to the large number of ironable icons, the model has higher accuracy in identifying this category; for non-ironable icons, although the number of sample data of this type is small , but this application will later detect the non-ironable icon recognized by the first recognition model through shape detection, thereby improving the accuracy of identifying the ironing mark.
在上述实施例的基础上，所述在确定衣物保养实施类型为可熨烫类型时，对所述水洗唛图像进行熨烫信息识别，并根据熨烫信息识别结果，获取对应的一种或多种熨烫参数识别方式，以通过所述熨烫参数识别方式，得到挂烫机的熨烫参数，包括：Based on the above embodiments, when it is determined that the clothing maintenance implementation type is an ironable type, ironing information recognition is performed on the washable mark image, and the corresponding one or more items are obtained based on the ironing information recognition result. An ironing parameter identification method to obtain the ironing parameters of the garment ironing machine through the ironing parameter identification method, including:
在确定所述衣物保养实施类型为可熨烫类型时，对所述水洗唛图像进行文字信息识别，获取对应的水洗唛文字内容识别结果；When it is determined that the clothing maintenance implementation type is the ironable type, text information recognition is performed on the wash mark image to obtain the corresponding wash mark text content recognition result;
若所述水洗唛文字内容识别结果中不包含任何文字信息，将所述水洗唛图像输入到第二识别模型中进行熨烫信息识别，得到由所述第二识别模型输出的熨烫图标信息，其中，所述第二识别模型是由熨烫图标标签的样本水洗唛图像训练得到的；If the recognition result of the washed mark text content does not contain any text information, input the washed mark image into the second recognition model to identify the ironing information, and obtain the ironing icon information output by the second recognition model, Wherein, the second recognition model is trained by sample washed mark images of ironing icon labels;
根据所述熨烫图标信息，生成所述挂烫机的熨烫参数。According to the ironing icon information, the ironing parameters of the garment ironing machine are generated.
在本申请中，可通过光学字符识别(Optical Character Recognition，OCR)技术，对第一识别模型识别得到的可熨烫图标对应区域进行文字信息识别，当该区域内不包含有文字内容的情况下，通过第二识别模型对可熨烫图标再一次进行识别，以得到熨烫图标信息，例如，熨烫图标中包括有2个“■”的符号，第二识别模型输出的结果为最高熨烫温度为150摄氏度，由于此类的熨烫图标的样本数量较多，使得第二识别模型基 于这些样本数据完成训练后，可精确的完成识别任务。In this application, optical character recognition (Optical Character Recognition, OCR) technology can be used to identify text information in the area corresponding to the ironable icon recognized by the first recognition model. When the area does not contain text content, , the ironable icon is recognized again through the second recognition model to obtain the ironing icon information. For example, the ironing icon includes two "■" symbols, and the output result of the second recognition model is the highest ironing The temperature is 150 degrees Celsius. Due to the large number of samples of this type of ironing icon, the second recognition model is basically After completing training on these sample data, the recognition task can be completed accurately.
在上述实施例的基础上，所述方法还包括：Based on the above embodiments, the method further includes:
若所述水洗唛文字内容识别结果中包含有文字信息，将所述水洗唛文字内容识别结果与预设熨烫规则进行类别映射，获取所述挂烫机对应的熨烫参数。If the recognition result of the washed mark text content contains text information, perform category mapping on the recognition result of the washed mark text content and the preset ironing rules to obtain the corresponding ironing parameters of the garment steamer.
在本申请中，根据OCR正则表达式的预设映射规则，进行类别映射，当水洗唛文字内容识别结果，通过预设熨烫规则可映射到对应类别时，获取熨烫参数，若无法映射，则表示该文字内容与熨烫不相关。In this application, category mapping is performed based on the preset mapping rules of OCR regular expressions. When the recognition result of the washed mark text content can be mapped to the corresponding category through the preset ironing rules, the ironing parameters are obtained. If the mapping cannot be performed, It means that the text content is not related to ironing.
在上述实施例的基础上，所述方法还包括：Based on the above embodiments, the method further includes:
在确定所述水洗唛文字内容识别结果中包含有文字信息的情况下，将所述水洗唛文字内容识别结果与预设熨烫规则进行类别映射，并将所述水洗唛图像输入到第二识别模型中进行熨烫信息识别；When it is determined that the recognition result of the text content of the washed mark contains text information, perform category mapping on the recognition result of the text content of the washed mark and the preset ironing rules, and input the image of the washed mark into the second recognition Ironing information recognition is carried out in the model;
基于类别映射结果和所述熨烫图标信息，获取所述挂烫机的熨烫参数。Based on the category mapping result and the ironing icon information, the ironing parameters of the garment ironing machine are obtained.
在本申请中，针对水洗唛图像中包含有文字信息的情况，可将两种熨烫参数识别方式进行组合，即将水洗唛文字内容识别结果与预设熨烫规则进行类别映射，同时将水洗唛图像输入到第二识别模型中进行熨烫信息识别，从而获取到两种熨烫参数识别方式各自对应的识别结果，最后再将两个识别结合进行合并，生成挂烫机对应的熨烫模式。In this application, for the case where the washed mark image contains text information, two ironing parameter identification methods can be combined, that is, the recognition results of the washed mark text content and the preset ironing rules are mapped to categories, and the washed mark image is The image is input into the second recognition model for ironing information recognition, thereby obtaining the recognition results corresponding to the two ironing parameter recognition methods. Finally, the two recognitions are combined and merged to generate the ironing mode corresponding to the garment ironing machine.
在上述实施例的基础上，在所述对水洗唛图像进行图像识别，基于图像识别结果进行目标检测之后，所述方法还包括：Based on the above embodiments, after performing image recognition on the water-washed mark image and performing target detection based on the image recognition results, the method further includes:
对所述水洗唛图像中水洗唛对应的区域范围进行图像裁剪处理，得到目标图像，以对所述目标图像进行熨烫信息识别或形状识别。An image cropping process is performed on the area range corresponding to the water-washed mark in the water-washed mark image to obtain a target image, so as to perform ironing information recognition or shape recognition on the target image.
在本申请中，当第一识别模型完成对水洗唛图像的初步识别后，可根据第一识别模型输出的结果，获取到水洗唛图像中熨烫图标的区域范围，进而通过对该区域范围进行裁剪，将仅包含有熨烫图标内容的图像作为目标图像，用于后续的目标识别检测，以减少识别过程中的干扰。In this application, after the first recognition model completes the preliminary recognition of the washed mark image, the area range of the ironing icon in the washed mark image can be obtained based on the results output by the first recognition model, and then the area range is processed Crop, and use the image containing only the ironing icon content as the target image for subsequent target recognition detection to reduce interference during the recognition process.
在一实施例中，对本申请的挂烫机熨烫方式进行整体说明，具体步骤如下：In one embodiment, an overall description of the garment ironing method of the present application is provided. The specific steps are as follows:
步骤101，将所有获取到的水洗熨烫图标训练数据分成两大类：可以 熨烫和不可熨烫(区分在于熨烫图标中间是否有个“×”符号)，然后按照目标检测方式进行标注，通过目标检测方法进行训练，得到用于识别可以熨烫/不可熨烫两类目标的第一识别模型。Step 101: Divide all obtained washing and ironing icon training data into two categories: OK Ironing and non-ironable (the difference lies in whether there is a "×" symbol in the middle of the ironing icon), and then marked according to the target detection method, and trained through the target detection method to obtain two categories for identifying ironable/non-ironable The first recognition model of the target.
步骤102，通过第一识别模型对水洗唛图像进行识别，如果识别结果为不可熨烫，至步骤103；如果识别结果为可以熨烫，至步骤106。Step 102: Recognize the washable label image through the first recognition model. If the recognition result is that it cannot be ironed, go to step 103; if the recognition result is that it can be ironed, go to step 106.
步骤103，把不可熨烫图标按照框坐标(x,y,w,h)裁剪为小图(即目标图像)，然后对小图进行传统的圆检测，三角形检测和矩形检测，如果有上述任何之一的形状存在，至步骤104；如果上述形状都没有，至步骤105。Step 103: Crop the non-ironable icon into a small image (i.e. target image) according to the frame coordinates (x, y, w, h), and then perform traditional circle detection, triangle detection and rectangle detection on the small image. If there are any of the above If one of the shapes exists, go to step 104; if none of the above shapes exists, go to step 105.
步骤104，若检出的小图为不可干洗，不可漂白或不可翻转干燥的图标，即小图中包含有圆形、三角形或矩形，则表示该图标不是不可熨烫图标，因此不输出该图标的信息，无需对挂烫机的熨烫模式进行调整。Step 104, if the detected thumbnail is an icon that cannot be dry cleaned, cannot be bleached or cannot be flipped and dried, that is, the thumbnail contains a circle, triangle or rectangle, it means that the icon is not a non-ironable icon, so the icon is not output. information, there is no need to adjust the ironing mode of the garment steamer.
步骤105，若检出的小图确实为不可熨烫图标，至步骤111。Step 105: If the detected thumbnail is indeed a non-ironable icon, go to step 111.
步骤106，把可熨烫图标按照框坐标(x,y,w,h)裁剪为小图，对小图进行OCR识别，会出现两种情况：OCR识别出文字(包括数字内容)，至步骤107；OCR识别不出任何文字，至步骤110。Step 106: Cut the ironable icon into a small picture according to the frame coordinates (x, y, w, h), and perform OCR recognition on the small picture. Two situations will occur: OCR recognizes text (including digital content), and then go to step 106. 107; OCR cannot recognize any text, go to step 110.
步骤107，根据OCR正则表达式的映射规则，进行类别映射，可能出现两种情况：可映射到类别，至步骤108；无法映射到对应类别，至步骤109。Step 107: Perform category mapping according to the mapping rules of the OCR regular expression. There may be two situations: if it can be mapped to a category, go to step 108; if it cannot be mapped to the corresponding category, go to step 109.
步骤108，根据映射得到的类别，生成对应的熨烫参数，比如文字内容为80-120摄氏度，可映射到80-120温度参数(为可输出的低频类别)，至步骤111。Step 108: Generate corresponding ironing parameters based on the mapped categories. For example, text content of 80-120 degrees Celsius can be mapped to 80-120 temperature parameters (which are low-frequency categories that can be output). Go to step 111.
步骤109，对于映射规则中不存在的类别，无需生成输出结果，比如，文字内容为“199”，通过正则对应不到任何类别，则不输出。Step 109: For categories that do not exist in the mapping rules, there is no need to generate output results. For example, if the text content is "199" and cannot be matched to any category through the regular expression, no output will be generated.
步骤110，当OCR识别不出任何文字时，直接通过第二识别模型对可熨烫图标进行识别，输出对应的熨烫参数。Step 110: When OCR cannot recognize any text, the ironable icon is directly recognized through the second recognition model and the corresponding ironing parameters are output.
步骤111，将步骤108和步骤110输出的结果进行合并，发送至挂烫机执行相应的熨烫模式。Step 111: Combine the output results of step 108 and step 110, and send them to the garment steamer to execute the corresponding ironing mode.
下面对本申请提供的挂烫机熨烫装置进行描述，下文描述的挂烫机熨烫装置与上文描述的挂烫机熨烫方法可相互对应参照。 The garment ironing device provided by the present application will be described below. The garment ironing device described below and the garment ironing method described above may be mutually referenced.
图3为本申请提供的挂烫机熨烫装置的结构示意图，如图3所示，本申请提供了一种挂烫机熨烫装置，包括衣物保养实施类型检测模块301、熨烫信息检测模块302和控制模块303，其中，衣物保养实施类型检测模块301用于对水洗唛图像进行图像识别，基于图像识别结果进行目标检测；熨烫信息检测模块302用于在确定衣物保养实施类型为可熨烫类型时，对所述水洗唛图像进行熨烫信息识别，并根据熨烫信息识别结果，获取对应的一种或多种熨烫参数识别方式，以通过所述熨烫参数识别方式，获取挂烫机的熨烫参数；控制模块303用于基于所述熨烫参数，生成所述挂烫机对应的熨烫模式，并根据所述熨烫模式控制所述挂烫机。Figure 3 is a schematic structural diagram of an ironing device for a garment steamer provided by this application. As shown in Figure 3, this application provides an ironing device for a garment steamer, including a clothing maintenance implementation type detection module 301 and an ironing information detection module. 302 and control module 303, wherein the clothing maintenance implementation type detection module 301 is used to perform image recognition on the washing mark image and perform target detection based on the image recognition results; the ironing information detection module 302 is used to determine that the clothing maintenance implementation type is ironable. In the case of ironing type, perform ironing information identification on the washed mark image, and obtain one or more corresponding ironing parameter identification methods according to the ironing information identification results, so as to obtain the hanger through the ironing parameter identification method. Ironing parameters of the ironing machine; the control module 303 is configured to generate an ironing mode corresponding to the garment ironing machine based on the ironing parameters, and control the garment ironing machine according to the ironing mode.
本申请提供的挂烫机熨烫装置，首先对水洗唛图像进行衣物保养方式识别，在确定衣物保养实施类型为可熨烫类型时，再对水洗唛图像进行熨烫信息识别，并根据熨烫信息识别结果，生成用于控制挂烫机的熨烫参数，相比现有技术，通过对水洗唛图像进行多阶段检测识别，避免目标检测模型由于训练样本过少导致识别效果较差的问题，提高了水洗唛图像中熨烫信息的识别准确率。The garment ironing device provided by this application first identifies the clothing maintenance method on the wash mark image. When it is determined that the clothing maintenance implementation type is an ironable type, it then identifies the ironing information on the wash mark image and performs the ironing process according to the ironing method. The information recognition results are used to generate ironing parameters for controlling the garment ironing machine. Compared with the existing technology, multi-stage detection and recognition of washed mark images is performed to avoid the problem of poor recognition effect of the target detection model due to too few training samples. Improved the recognition accuracy of ironing information in washed mark images.
在上述实施例的基础上，所述装置还包括水洗唛形状检测模块和衣物保养提示模块，其中，水洗唛形状检测模块用于在确定所述衣物保养实施类型为不可熨烫类型时，对所述水洗唛图像中的水洗唛进行形状识别，并根据形状识别结果，确定所述水洗唛图像中的水洗唛是否为不可熨烫图标；衣物保养提示模块用于若所述水洗唛不是不可熨烫图标，根据所述水洗唛的形状识别结果，确定所述水洗唛对应的衣物保养不可实施信息，并通过所述挂烫机展示所述衣物保养不可实施信息，其中，所述衣物保养不可实施信息包括不可洗涤、不可漂白和不可翻转干燥中的一种或多种。On the basis of the above embodiment, the device further includes a washing mark shape detection module and a clothing care prompt module, wherein the washing mark shape detection module is used to determine whether the clothing maintenance implementation type is a non-ironable type. The wash mark in the wash mark image is subjected to shape recognition, and based on the shape recognition result, it is determined whether the wash mark in the wash mark image is a non-iron icon; the clothing care prompt module is used to determine if the wash mark is not a non-iron icon. Icon, according to the shape recognition result of the wash mark, determine the clothing maintenance unimplementable information corresponding to the wash mark, and display the clothing maintenance unimplementable information through the garment steamer, wherein the clothing maintenance unimplementable information Including one or more of do not wash, do not bleach and do not tumble dry.
在上述实施例的基础上，所述水洗唛形状检测模块具体用于，若所述水洗唛图像中不包括任何预设水洗唛形状时，确定所述水洗唛图像中的水洗唛为不可熨烫图标，其中，所述预设水洗唛形状包括圆形、三角形和矩形中的一种或几种的组合。Based on the above embodiments, the washable mark shape detection module is specifically used to determine that the washable mark in the washable mark image is non-ironable if the washable mark image does not include any preset washable mark shape. icon, wherein the preset washable mark shape includes one or a combination of circles, triangles and rectangles.
在上述实施例的基础上，所述衣物保养实施类型检测模块包括图像 采集单元和第一识别单元，其中，图像采集单元用于获取目标衣物的水洗唛图像；第一识别单元用于将所述水洗唛图像输入到第一识别模型中进行衣物保养方式识别，得到由所述第一识别模型输出的衣物保养实施类型，其中，所述第一识别模型是由标记有衣物保养实施类型标签的样本水洗唛图像训练得到的。Based on the above embodiments, the clothing maintenance implementation type detection module includes an image The acquisition unit and the first recognition unit, wherein the image acquisition unit is used to acquire the wash mark image of the target clothing; the first recognition unit is used to input the wash mark image into the first recognition model to identify the clothing maintenance method, and obtain the The clothing care implementation type output by the first recognition model, wherein the first recognition model is trained by sample washing mark images marked with clothing care implementation type labels.
在上述实施例的基础上，所述装置还包括样本图像获取模块、训练集构建模块和训练模块，其中，样本图像获取模块用于获取多个样本水洗唛图像；训练集构建模块用于为各个所述样本水洗唛图像标记对应的衣物保养实施类型标签，构建得到训练样本集，其中，所述衣物保养实施类型标签至少包括可熨烫标签和不可熨烫标签；训练模块用于通过所述训练样本集，对卷积神经网络进行训练，得到所述第一识别模型。On the basis of the above embodiments, the device further includes a sample image acquisition module, a training set construction module and a training module, wherein the sample image acquisition module is used to obtain multiple sample water-washed mark images; the training set construction module is used to provide each The sample washing mark image is marked with a clothing care implementation type label corresponding to a training sample set, wherein the clothing care implementation type label at least includes an ironable label and a non-ironable label; the training module is used to pass the training The sample set is used to train the convolutional neural network to obtain the first recognition model.
在上述实施例的基础上，所述熨烫信息检测模块包括文字识别单元、第二识别单元和熨烫参数生成单元，其中，文字识别单元用于在确定所述衣物保养实施类型为可熨烫类型时，对所述水洗唛图像进行文字信息识别，获取对应的水洗唛文字内容识别结果；第二识别单元用于若所述水洗唛文字内容识别结果中不包含任何文字信息，将所述水洗唛图像输入到第二识别模型中进行熨烫信息识别，得到由所述第二识别模型输出的熨烫图标信息，其中，所述第二识别模型是由熨烫图标标签的样本水洗唛图像训练得到的；熨烫参数生成单元用于根据所述熨烫图标信息，生成所述挂烫机的熨烫参数。Based on the above embodiments, the ironing information detection module includes a text recognition unit, a second recognition unit and an ironing parameter generation unit, wherein the text recognition unit is used to determine that the clothing maintenance implementation type is ironable. type, perform text information recognition on the water-washed mark image, and obtain the corresponding water-washed mark text content recognition result; the second recognition unit is used to identify the water-washed mark text content if the water-washed mark text content recognition result does not contain any text information. The mark image is input into the second recognition model for ironing information recognition, and the ironing icon information output by the second recognition model is obtained, wherein the second recognition model is trained by sample washed mark images of the ironing icon label. Obtained; the ironing parameter generating unit is used to generate the ironing parameters of the garment ironing machine according to the ironing icon information.
在上述实施例的基础上，所述装置还包括熨烫参数映射模块，用于若所述水洗唛文字内容识别结果中包含有文字信息，将所述水洗唛文字内容识别结果与预设熨烫规则进行类别映射，获取所述挂烫机的熨烫参数。Based on the above embodiment, the device further includes an ironing parameter mapping module, which is used to, if the recognition result of the text content of the washed mark contains text information, compare the recognition result of the text content of the washed mark with the preset ironing The rules perform category mapping to obtain the ironing parameters of the garment ironing machine.
在上述实施例的基础上，所述熨烫信息检测模块还用于：在确定所述水洗唛文字内容识别结果中包含有文字信息的情况下，将所述水洗唛文字内容识别结果与预设熨烫规则进行类别映射，并将所述水洗唛图像输入到第二识别模型中进行熨烫信息识别；基于类别映射结果和所述熨烫图标信息，获取所述挂烫机的熨烫参数。Based on the above embodiment, the ironing information detection module is also used to: when it is determined that the text content recognition result of the water-washed mark contains text information, compare the recognition result of the text content of the water-washed mark with the preset The ironing rules perform category mapping, and the washed mark image is input into the second recognition model to identify the ironing information; based on the category mapping results and the ironing icon information, the ironing parameters of the garment ironing machine are obtained.
在上述实施例的基础上，所述装置还包括图像采集模块，用于对所 述水洗唛图像中水洗唛对应的区域范围进行图像裁剪处理，得到目标图像，以对所述目标图像进行熨烫信息识别或形状识别。Based on the above embodiments, the device further includes an image acquisition module for The area range corresponding to the water-washed mark in the water-washed mark image is subjected to image cropping processing to obtain a target image, so as to perform ironing information recognition or shape recognition on the target image.
图5为本申请提供的挂烫机的结构示意图。如图5所示，本申请还提供一种挂烫机，包括上述各实施例所述的挂烫机熨烫装置，以及蒸汽发热器、机身外壳、水箱、蒸汽喷头、蒸汽导管和可伸缩支架。本申请提供的挂烫机，通过挂烫机熨烫装置，对水洗唛图像进行衣物保养方式识别，在确定衣物保养实施类型为可熨烫类型时，再对水洗唛图像进行熨烫信息识别，并根据熨烫信息识别结果，生成用于控制挂烫机的熨烫参数，相比现有技术，通过对水洗唛图像进行多阶段检测识别，避免目标检测模型由于训练样本过少导致识别效果较差的问题，提高了水洗唛图像中熨烫信息的识别准确率。Figure 5 is a schematic structural diagram of the garment steamer provided by this application. As shown in Figure 5, this application also provides a garment steamer, including the garment steamer ironing device described in the above embodiments, as well as a steam heater, a body shell, a water tank, a steam nozzle, a steam conduit and a retractable stand. The garment ironing machine provided by this application uses the ironing device of the garment ironing machine to identify the clothing maintenance method on the washing mark image. When it is determined that the clothing maintenance implementation type is an ironable type, the washing mark image is then identified as ironing information. And based on the ironing information recognition results, ironing parameters for controlling the garment ironing machine are generated. Compared with the existing technology, multi-stage detection and recognition of washed mark images is performed to avoid poor recognition results of the target detection model due to too few training samples. It solves the problem of poor quality and improves the recognition accuracy of ironing information in washed mark images.
图4为本申请提供的电子装置的结构示意图，如图4所示，该电子装置可以包括：处理器(processor)410、通信接口(Communications Interface)420、存储器(memory)430和通信总线440，其中，处理器410，通信接口420，存储器430通过通信总线440完成相互间的通信。处理器410可以调用存储器430中的逻辑指令，以执行挂烫机熨烫方法，该方法包括：对水洗唛图像进行图像识别，基于图像识别结果进行目标检测；在确定衣物保养实施类型为可熨烫类型时，对所述水洗唛图像进行熨烫信息识别，并根据熨烫信息识别结果，获取对应的一种或多种熨烫参数识别方式，以通过所述熨烫参数识别方式，得到挂烫机的熨烫参数；基于所述熨烫参数，生成所述挂烫机对应的熨烫模式，并根据所述熨烫模式控制所述挂烫机。Figure 4 is a schematic structural diagram of an electronic device provided by this application. As shown in Figure 4, the electronic device may include: a processor (processor) 410, a communications interface (Communications Interface) 420, a memory (memory) 430 and a communication bus 440. Among them, the processor 410, the communication interface 420, and the memory 430 complete communication with each other through the communication bus 440. The processor 410 can call the logical instructions in the memory 430 to execute the garment ironing method. The method includes: performing image recognition on the wash mark image, performing target detection based on the image recognition result; and determining that the clothing maintenance implementation type is ironable. In the case of ironing type, perform ironing information identification on the washed mark image, and obtain one or more corresponding ironing parameter identification methods according to the ironing information identification results, so as to obtain hanging information through the ironing parameter identification method. Ironing parameters of the ironing machine; based on the ironing parameters, generate an ironing mode corresponding to the garment ironing machine, and control the garment ironing machine according to the ironing mode.
此外，上述的存储器430中的逻辑指令可以通过软件功能单元的形式实现并作为独立的产品销售或使用时，可以存储在一个计算机可读取存储介质中。基于这样的理解，本申请的技术方案本质上或者说对现有技术做出贡献的部分或者该技术方案的部分可以以软件产品的形式体现出来，该计算机软件产品存储在一个存储介质中，包括若干指令用以使得一台计算机设备(可以是个人计算机，服务器，或者网络设备等)执行本申请各个实施例所述方法的全部或部分步骤。而前述的存储介质包括：U盘、移动硬盘、只读存储器(Read-Only Memory，ROM)、随机 存取存储器(Random Access Memory，RAM)、磁碟或者光盘等各种可以存储程序代码的介质。In addition, the above-mentioned logical instructions in the memory 430 can be implemented in the form of software functional units and can be stored in a computer-readable storage medium when sold or used as an independent product. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the existing technology or the part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including Several instructions are used to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of this application. The aforementioned storage media include: U disk, mobile hard disk, read-only memory (Read-Only Memory, ROM), random Random Access Memory (RAM), magnetic disks, optical disks and other media that can store program code.
另一方面，本申请还提供一种计算机程序产品，所述计算机程序产品包括计算机程序，计算机程序可存储在计算机可读的存储介质上，所述计算机程序被处理器执行时，计算机能够执行上述各方法所提供的挂烫机熨烫方法，该方法包括：对水洗唛图像进行图像识别，基于图像识别结果进行目标检测；在确定衣物保养实施类型为可熨烫类型时，对所述水洗唛图像进行熨烫信息识别，并根据熨烫信息识别结果，获取对应的一种或多种熨烫参数识别方式，以通过所述熨烫参数识别方式，得到挂烫机的熨烫参数；基于所述熨烫参数，生成所述挂烫机对应的熨烫模式，并根据所述熨烫模式控制所述挂烫机。On the other hand, this application also provides a computer program product. The computer program product includes a computer program. The computer program can be stored on a computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the above-mentioned steps. The garment ironing method provided by each method includes: performing image recognition on the washable mark image, and performing target detection based on the image recognition result; when it is determined that the clothing maintenance implementation type is an ironable type, performing image recognition on the washable mark image. The image is used to identify ironing information, and according to the ironing information identification result, one or more corresponding ironing parameter identification methods are obtained, so as to obtain the ironing parameters of the garment ironing machine through the ironing parameter identification method; based on the The ironing parameters are used to generate an ironing mode corresponding to the garment ironing machine, and the garment ironing machine is controlled according to the ironing mode.
又一方面，本申请还提供一种计算机可读的存储介质，所述计算机可读的存储介质包括存储的程序，其中，所述程序运行时执行上述各方法提供的挂烫机熨烫方法，该方法包括：对水洗唛图像进行图像识别，基于图像识别结果进行目标检测；在确定衣物保养实施类型为可熨烫类型时，对所述水洗唛图像进行熨烫信息识别，并根据熨烫信息识别结果，获取对应的一种或多种熨烫参数识别方式，以通过所述熨烫参数识别方式，得到挂烫机的熨烫参数；基于所述熨烫参数，生成所述挂烫机对应的熨烫模式，并根据所述熨烫模式控制所述挂烫机。In another aspect, the present application also provides a computer-readable storage medium, which includes a stored program, wherein when the program is run, the garment ironing method provided by each of the above methods is executed, The method includes: performing image recognition on the wash mark image, and performing target detection based on the image recognition result; when determining that the clothing maintenance implementation type is an ironable type, performing ironing information recognition on the wash mark image, and performing ironing information identification on the wash mark image based on the ironing information. Recognition results, obtain the corresponding one or more ironing parameter identification methods, so as to obtain the ironing parameters of the garment ironing machine through the ironing parameter identification method; based on the ironing parameters, generate the corresponding information of the garment ironing machine an ironing mode, and control the garment ironing machine according to the ironing mode.
以上所描述的装置实施例仅仅是示意性的，其中所述作为分离部件说明的单元可以是或者也可以不是物理上分开的，作为单元显示的部件可以是或者也可以不是物理单元，即可以位于一个地方，或者也可以分布到多个网络单元上。可以根据实际的需要选择其中的部分或者全部模块来实现本实施例方案的目的。本领域普通技术人员在不付出创造性的劳动的情况下，即可以理解并实施。The device embodiments described above are only illustrative. The units described as separate components may or may not be physically separated. The components shown as units may or may not be physical units, that is, they may be located in One location, or it can be distributed across multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. Persons of ordinary skill in the art can understand and implement the method without any creative effort.
通过以上的实施方式的描述，本领域的技术人员可以清楚地了解到各实施方式可借助软件加必需的通用硬件平台的方式来实现，当然也可以通过硬件。基于这样的理解，上述技术方案本质上或者说对现有技术做出贡献的部分可以以软件产品的形式体现出来，该计算机软件产品可以存储在计算机可读存储介质中，如ROM/RAM、磁碟、光盘等，包括 若干指令用以使得一台计算机设备(可以是个人计算机，服务器，或者网络设备等)执行各个实施例或者实施例的某些部分所述的方法。Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the part of the above technical solution that essentially contributes to the existing technology can be embodied in the form of a software product. The computer software product can be stored in a computer-readable storage medium, such as ROM/RAM, magnetic Discs, compact discs, etc., including Several instructions are used to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in various embodiments or certain parts of the embodiments.
最后应说明的是：以上实施例仅用以说明本申请的技术方案，而非对其限制；尽管参照前述实施例对本申请进行了详细的说明，本领域的普通技术人员应当理解：其依然可以对前述各实施例所记载的技术方案进行修改，或者对其中部分技术特征进行等同替换；而这些修改或者替换，并不使相应技术方案的本质脱离本申请各实施例技术方案的精神和范围。 Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present application, but not to limit it; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that it can still be Modifications are made to the technical solutions described in the foregoing embodiments, or equivalent substitutions are made to some of the technical features; however, these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions in the embodiments of the present application.

Claims (21)

  1. 一种挂烫机熨烫方法，包括：An ironing method using a garment steamer, including:
    对水洗唛图像进行图像识别，基于图像识别结果进行目标检测；Perform image recognition on washed mark images and perform target detection based on the image recognition results;
    在确定衣物保养实施类型为可熨烫类型时，对所述水洗唛图像进行熨烫信息识别，并根据熨烫信息识别结果，获取对应的一种或多种熨烫参数识别方式，以通过所述熨烫参数识别方式，得到挂烫机的熨烫参数；When it is determined that the clothing maintenance implementation type is an ironable type, ironing information recognition is performed on the washable mark image, and the corresponding one or more ironing parameter recognition methods are obtained based on the ironing information recognition results to pass the required Use the ironing parameter identification method described above to obtain the ironing parameters of the garment ironing machine;
    基于所述熨烫参数，生成所述挂烫机对应的熨烫模式，并根据所述熨烫模式控制所述挂烫机。Based on the ironing parameters, an ironing mode corresponding to the garment ironing machine is generated, and the garment ironing machine is controlled according to the ironing mode.
  2. 根据权利要求1所述的挂烫机熨烫方法，还包括：The garment ironing method according to claim 1, further comprising:
    在确定所述衣物保养实施类型为不可熨烫类型时，对所述水洗唛图像中的水洗唛进行形状识别，并根据形状识别结果，确定所述水洗唛图像中的水洗唛是否为不可熨烫图标；When it is determined that the clothing maintenance implementation type is a non-ironable type, shape recognition is performed on the washable mark in the washable mark image, and based on the shape recognition result, it is determined whether the washable mark in the washable mark image is non-ironable. icon;
    若所述水洗唛不是不可熨烫图标，根据所述水洗唛的形状识别结果，确定所述水洗唛对应的衣物保养不可实施信息，并通过所述挂烫机展示所述衣物保养不可实施信息，其中，所述衣物保养不可实施信息包括不可洗涤、不可漂白和不可翻转干燥中的一种或多种。If the washable mark is not a non-ironable icon, determine the clothing maintenance disabling information corresponding to the washable mark based on the shape recognition result of the washable mark, and display the clothing maintenance disabling information through the garment ironing machine, Wherein, the clothing maintenance disabling information includes one or more of the following: no washing, no bleaching, and no flipping and drying.
  3. 根据权利要求2所述的挂烫机熨烫方法，其中，所述根据形状识别结果，确定所述水洗唛图像中的水洗唛是否为不可熨烫图标，包括：The ironing method of a garment steamer according to claim 2, wherein determining whether the washable mark in the washable mark image is a non-ironable icon according to the shape recognition result includes:
    若所述水洗唛图像中不包括任何预设水洗唛形状时，确定所述水洗唛图像中的水洗唛为不可熨烫图标，其中，所述预设水洗唛形状包括圆形、三角形和矩形中的一种或几种的组合。If the washable mark image does not include any preset washable mark shape, it is determined that the washable mark in the washable mark image is a non-ironable icon, wherein the preset washable mark shapes include circles, triangles and rectangles. one or a combination of several.
  4. 根据权利要求1所述的挂烫机熨烫方法，其中，所述对水洗唛图像进行图像识别，包括：The garment ironing method according to claim 1, wherein the image recognition of the washed mark image includes:
    获取目标衣物的水洗唛图像；Obtain the washing mark image of the target clothing;
    将所述水洗唛图像输入到第一识别模型中进行衣物保养方式识别，得到由所述第一识别模型输出的衣物保养实施类型，其中，所述第一识别模型是由标记有衣物保养实施类型标签的样本水洗唛图像训练得到的。Input the washed mark image into the first recognition model to identify the clothing care mode, and obtain the clothing care implementation type output by the first recognition model, wherein the first recognition model is a clothing care implementation type marked with The labeled sample water-washed label images are trained.
  5. 根据权利要求4所述的挂烫机熨烫方法，其中，所述第一识别模 型通过以下步骤训练得到：The ironing method of a garment steamer according to claim 4, wherein the first identification pattern The model is trained through the following steps:
    获取多个样本水洗唛图像；Obtain water-washed mark images of multiple samples;
    为各个所述样本水洗唛图像标记对应的衣物保养实施类型标签，构建得到训练样本集，其中，所述衣物保养实施类型标签至少包括可熨烫标签和不可熨烫标签；Label each sample washing mark image with a corresponding clothing care implementation type label to construct a training sample set, wherein the clothing care implementation type label at least includes an ironable label and a non-ironable label;
    通过所述训练样本集，对卷积神经网络进行训练，得到所述第一识别模型。The convolutional neural network is trained through the training sample set to obtain the first recognition model.
  6. 根据权利要求1所述的挂烫机熨烫方法，其中，所述在确定衣物保养实施类型为可熨烫类型时，对所述水洗唛图像进行熨烫信息识别，并根据熨烫信息识别结果，获取对应的一种或多种熨烫参数识别方式，以通过所述熨烫参数识别方式，得到挂烫机的熨烫参数，包括：The ironing method of a garment ironing machine according to claim 1, wherein when it is determined that the clothing maintenance implementation type is an ironable type, ironing information recognition is performed on the washable mark image, and the ironing information recognition result is , obtain the corresponding one or more ironing parameter identification methods, so as to obtain the ironing parameters of the garment ironing machine through the ironing parameter identification methods, including:
    在确定所述衣物保养实施类型为可熨烫类型时，对所述水洗唛图像进行文字信息识别，获取对应的水洗唛文字内容识别结果；When it is determined that the clothing maintenance implementation type is the ironable type, text information recognition is performed on the wash mark image to obtain the corresponding wash mark text content recognition result;
    若所述水洗唛文字内容识别结果中不包含任何文字信息，将所述水洗唛图像输入到第二识别模型中进行熨烫信息识别，得到由所述第二识别模型输出的熨烫图标信息，其中，所述第二识别模型是由标记有熨烫图标标签的样本水洗唛图像训练得到的；If the recognition result of the washed mark text content does not contain any text information, input the washed mark image into the second recognition model to identify the ironing information, and obtain the ironing icon information output by the second recognition model, Wherein, the second recognition model is trained by sample washed mark images marked with ironing icon labels;
    根据所述熨烫图标信息，生成所述挂烫机的熨烫参数。According to the ironing icon information, the ironing parameters of the garment ironing machine are generated.
  7. 根据权利要求6所述的挂烫机熨烫方法，还包括：The garment ironing method according to claim 6, further comprising:
    若所述水洗唛文字内容识别结果中包含有文字信息，将所述水洗唛文字内容识别结果与预设熨烫规则进行类别映射，获取所述挂烫机的熨烫参数。If the recognition result of the washed mark text content contains text information, perform category mapping on the recognition result of the washed mark text content and the preset ironing rules to obtain the ironing parameters of the garment ironing machine.
  8. 根据权利要求7所述的挂烫机熨烫方法，还包括：The garment ironing method according to claim 7, further comprising:
    在确定所述水洗唛文字内容识别结果中包含有文字信息的情况下，将所述水洗唛文字内容识别结果与预设熨烫规则进行类别映射，并将所述水洗唛图像输入到第二识别模型中进行熨烫信息识别；When it is determined that the recognition result of the text content of the washed mark contains text information, perform category mapping on the recognition result of the text content of the washed mark and the preset ironing rules, and input the image of the washed mark into the second recognition Ironing information recognition is carried out in the model;
    基于类别映射结果和所述熨烫图标信息，获取所述挂烫机的熨烫参数。Based on the category mapping result and the ironing icon information, the ironing parameters of the garment ironing machine are obtained.
  9. 根据权利要求2至8任一项所述的挂烫机熨烫方法，其中，在所述对水洗唛图像进行图像识别，基于图像识别结果进行目标检测之后， 所述方法还包括：The garment ironing method according to any one of claims 2 to 8, wherein after performing image recognition on the washed mark image and performing target detection based on the image recognition result, The method also includes:
    对所述水洗唛图像中水洗唛对应的区域范围进行图像裁剪处理，得到目标图像，以对所述目标图像进行熨烫信息识别或形状识别。An image cropping process is performed on the area range corresponding to the water-washed mark in the water-washed mark image to obtain a target image, so as to perform ironing information recognition or shape recognition on the target image.
  10. 一种挂烫机熨烫装置，包括：An ironing device for a garment steamer, including:
    衣物保养实施类型检测模块，设置为对水洗唛图像进行图像识别，基于图像识别结果进行目标检测；The clothing maintenance implementation type detection module is set to perform image recognition on the washed mark image and perform target detection based on the image recognition results;
    熨烫信息检测模块，设置为在确定衣物保养实施类型为可熨烫类型时，对所述水洗唛图像进行熨烫信息识别，并根据熨烫信息识别结果，获取对应的一种或多种熨烫参数识别方式，以通过所述熨烫参数识别方式，获取挂烫机的熨烫参数；The ironing information detection module is configured to perform ironing information identification on the washed mark image when it is determined that the clothing maintenance implementation type is an ironable type, and obtain the corresponding one or more ironing information based on the ironing information identification results. An ironing parameter identification method to obtain the ironing parameters of the garment ironing machine through the ironing parameter identification method;
    控制模块，设置为基于所述熨烫参数，生成所述挂烫机对应的熨烫模式，并根据所述熨烫模式控制所述挂烫机。A control module is configured to generate an ironing mode corresponding to the garment ironing machine based on the ironing parameters, and control the garment ironing machine according to the ironing mode.
  11. 根据权利要求10所述的挂烫机熨烫装置，还包括：The garment ironing device according to claim 10, further comprising:
    水洗唛形状检测模块，设置为在确定所述衣物保养实施类型为不可熨烫类型时，对所述水洗唛图像中的水洗唛进行形状识别，并根据形状识别结果，确定所述水洗唛图像中的水洗唛是否为不可熨烫图标；The wash mark shape detection module is configured to perform shape recognition on the wash mark in the wash mark image when it is determined that the clothing maintenance implementation type is a non-ironing type, and determine the shape of the wash mark in the wash mark image based on the shape recognition result. Whether the washable mark is a non-ironable icon;
    衣物保养提示模块，设置为若所述水洗唛不是不可熨烫图标，根据所述水洗唛的形状识别结果，确定所述水洗唛对应的衣物保养不可实施信息，并通过所述挂烫机展示所述衣物保养不可实施信息，其中，所述衣物保养不可实施信息包括不可洗涤、不可漂白和不可翻转干燥中的一种或多种。The clothing care prompt module is configured to determine the clothing maintenance unimplementable information corresponding to the wash mark based on the shape recognition result of the wash mark if the wash mark is not a non-ironing icon, and display the information through the garment ironing machine. The clothing maintenance disabling information includes one or more of the following: no washing, no bleaching, and no flipping and drying.
  12. 根据权利要求11所述的挂烫机熨烫装置，所述水洗唛形状检测模块具体用于：According to the garment ironing device of claim 11, the washable mark shape detection module is specifically used for:
    若所述水洗唛图像中不包括任何预设水洗唛形状时，确定所述水洗唛图像中的水洗唛为不可熨烫图标，其中，所述预设水洗唛形状包括圆形、三角形和矩形中的一种或几种的组合。If the washable mark image does not include any preset washable mark shape, it is determined that the washable mark in the washable mark image is a non-ironable icon, wherein the preset washable mark shapes include circles, triangles and rectangles. one or a combination of several.
  13. 根据权利要求10所述的挂烫机熨烫装置，其中，所述衣物保养实施类型检测模块包括：The garment ironing device according to claim 10, wherein the clothing maintenance implementation type detection module includes:
    图像采集单元，设置为获取目标衣物的水洗唛图像；The image acquisition unit is configured to acquire the washing mark image of the target clothing;
    第一识别单元，设置为将所述水洗唛图像输入到第一识别模型中进 行衣物保养方式识别，得到由所述第一识别模型输出的衣物保养实施类型，其中，所述第一识别模型是由标记有衣物保养实施类型标签的样本水洗唛图像训练得到的。The first recognition unit is configured to input the water-washed mark image into the first recognition model for The clothing care method is recognized to obtain the clothing care implementation type output by the first recognition model, wherein the first recognition model is trained by sample washing mark images marked with clothing care implementation type labels.
  14. 根据权利要求13所述的挂烫机熨烫装置，其中，所述第一识别模型通过以下步骤训练得到：The garment ironing device according to claim 13, wherein the first recognition model is trained through the following steps:
    获取多个样本水洗唛图像；Obtain multiple sample water-washed mark images;
    为各个所述样本水洗唛图像标记对应的衣物保养实施类型标签，构建得到训练样本集，其中，所述衣物保养实施类型标签至少包括可熨烫标签和不可熨烫标签；Label each sample washing mark image with a corresponding clothing care implementation type label to construct a training sample set, wherein the clothing care implementation type label at least includes an ironable label and a non-ironable label;
    通过所述训练样本集，对卷积神经网络进行训练，得到所述第一识别模型。The convolutional neural network is trained through the training sample set to obtain the first recognition model.
  15. 根据权利要求10所述的挂烫机熨烫装置，其中，所述熨烫信息检测模块包括：The ironing device of a garment steamer according to claim 10, wherein the ironing information detection module includes:
    文字识别单元，设置为在确定所述衣物保养实施类型为可熨烫类型时，对所述水洗唛图像进行文字信息识别，获取对应的水洗唛文字内容识别结果；A text recognition unit, configured to perform text information recognition on the washed mark image when it is determined that the clothing maintenance implementation type is an ironable type, and obtain the corresponding wash mark text content recognition result;
    第二识别单元，设置为若所述水洗唛文字内容识别结果中不包含任何文字信息，将所述水洗唛图像输入到第二识别模型中进行熨烫信息识别，得到由所述第二识别模型输出的熨烫图标信息，其中，所述第二识别模型是由标记有熨烫图标标签的样本水洗唛图像训练得到的；The second recognition unit is configured to input the washed mark image into the second recognition model to identify the ironing information if the recognition result of the washed mark text content does not contain any text information, and obtain the second recognition model The output ironing icon information, wherein the second recognition model is trained by sample washed mark images marked with ironing icon labels;
    熨烫参数生成单元，设置为根据所述熨烫图标信息，生成所述挂烫机的熨烫参数。An ironing parameter generating unit is configured to generate ironing parameters of the garment ironing machine based on the ironing icon information.
  16. 根据权利要求15所述的挂烫机熨烫装置，还包括：The garment ironing device according to claim 15, further comprising:
    熨烫参数映射模块，设置为若所述水洗唛文字内容识别结果中包含有文字信息，将所述水洗唛文字内容识别结果与预设熨烫规则进行类别映射，获取所述挂烫机的熨烫参数。The ironing parameter mapping module is configured to perform category mapping between the recognition result of the washed mark text content and the preset ironing rules if the recognition result of the washed mark text content contains text information, and obtain the ironing function of the garment ironing machine. Hot parameters.
  17. 根据权利要求16所述的挂烫机熨烫装置，所述熨烫信息检测模块还用于：According to the ironing device of a garment ironing machine according to claim 16, the ironing information detection module is also used for:
    在确定所述水洗唛文字内容识别结果中包含有文字信息的情况下，将所述水洗唛文字内容识别结果与预设熨烫规则进行类别映射，并将所 述水洗唛图像输入到第二识别模型中进行熨烫信息识别；When it is determined that the recognition result of the text content of the washed mark contains text information, perform category mapping on the recognition result of the text content of the washed mark and the preset ironing rules, and The water-washed mark image is input into the second recognition model for ironing information recognition;
    基于类别映射结果和所述熨烫图标信息，获取所述挂烫机的熨烫参数。Based on the category mapping result and the ironing icon information, the ironing parameters of the garment ironing machine are obtained.
  18. 根据权利要求11至17任一项所述的挂烫机熨烫装置，还包括：The garment ironing device according to any one of claims 11 to 17, further comprising:
    图像采集模块，设置为对所述水洗唛图像中水洗唛对应的区域范围进行图像裁剪处理，得到目标图像，以对所述目标图像进行熨烫信息识别或形状识别。The image acquisition module is configured to perform image cropping processing on the area range corresponding to the water-washed mark in the water-washed mark image to obtain a target image, so as to perform ironing information recognition or shape recognition on the target image.
  19. 一种挂烫机，包括权利要求10至18任一项所述的挂烫机熨烫装置，以及蒸汽发热器、机身外壳、水箱、蒸汽喷头、蒸汽导管和可伸缩支架。A garment steamer, including the garment steamer ironing device according to any one of claims 10 to 18, as well as a steam heater, a body shell, a water tank, a steam nozzle, a steam conduit and a retractable bracket.
  20. 一种计算机可读的存储介质，所述计算机可读的存储介质包括存储的程序，其中，所述程序运行时执行权利要求1至9中任一项所述的方法。A computer-readable storage medium includes a stored program, wherein the method of any one of claims 1 to 9 is executed when the program is run.
  21. 一种电子装置，包括存储器和处理器，其中，所述存储器中存储有计算机程序，所述处理器被设置为通过所述计算机程序执行权利要求1至9中任一项所述的方法。 An electronic device includes a memory and a processor, wherein a computer program is stored in the memory, and the processor is configured to execute the method according to any one of claims 1 to 9 through the computer program.
PCT/CN2023/085376 2022-08-30 2023-03-31 Ironing method and device for garment steamer, and garment steamer, storage medium and electronic device WO2024045601A1 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202211049158.9 2022-08-30
CN202211049158.9A CN117689920A (en) 2022-08-30 2022-08-30 Garment steamer ironing method and device, garment steamer, storage medium and electronic device

Publications (1)

Publication Number Publication Date
WO2024045601A1 true WO2024045601A1 (en) 2024-03-07

Family

ID=90100293

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/085376 WO2024045601A1 (en) 2022-08-30 2023-03-31 Ironing method and device for garment steamer, and garment steamer, storage medium and electronic device

Country Status (2)

Country Link
CN (1) CN117689920A (en)
WO (1) WO2024045601A1 (en)

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106283584A (en) * 2016-10-28 2017-01-04 京东方科技集团股份有限公司 A kind of electric iron and ironing method thereof
KR102079661B1 (en) * 2019-05-24 2020-02-20 남현욱 Intelligence technique steam iron
CN112252001A (en) * 2020-08-31 2021-01-22 青岛云裳羽衣物联科技有限公司 Intelligent clothes nursing mirror with RFID function
CN113417107A (en) * 2021-06-30 2021-09-21 青岛海尔科技有限公司 Method and system for determining washing mode, storage medium and electronic device
CN114481584A (en) * 2022-01-17 2022-05-13 青岛海尔智护电器有限公司 Ironing method of garment steamer, control device, garment steamer and electronic equipment
CN114763671A (en) * 2021-01-15 2022-07-19 广州奕至家居科技有限公司 Ironing method based on clothes identification, terminal, ironing machine and storage medium

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106283584A (en) * 2016-10-28 2017-01-04 京东方科技集团股份有限公司 A kind of electric iron and ironing method thereof
KR102079661B1 (en) * 2019-05-24 2020-02-20 남현욱 Intelligence technique steam iron
CN112252001A (en) * 2020-08-31 2021-01-22 青岛云裳羽衣物联科技有限公司 Intelligent clothes nursing mirror with RFID function
CN114763671A (en) * 2021-01-15 2022-07-19 广州奕至家居科技有限公司 Ironing method based on clothes identification, terminal, ironing machine and storage medium
CN113417107A (en) * 2021-06-30 2021-09-21 青岛海尔科技有限公司 Method and system for determining washing mode, storage medium and electronic device
CN114481584A (en) * 2022-01-17 2022-05-13 青岛海尔智护电器有限公司 Ironing method of garment steamer, control device, garment steamer and electronic equipment

Also Published As

Publication number Publication date
CN117689920A (en) 2024-03-12

Similar Documents

Publication Publication Date Title
CN108914489B (en) Control method and device of washing machine, washing machine and storage medium
CN108411565B (en) Washing machine control method and device
JP7239782B2 (en) Tuning instrument settings with multi-pass training of target detection models
CN107904860B (en) Method and device for treating clothes in washing machine
WO2018192110A1 (en) Clothing care machine control method and apparatus, clothing care machine, and storage medium
CN108130683A (en) A kind of household electrical appliance and the method for running household electrical appliance
CN109112774B (en) Control method and device of washing machine, storage medium and washing machine
WO2021093359A1 (en) Method and system for controlling machines based on object recognition
CN107862313B (en) Dish washing machine and control method and device thereof
CN109355863B (en) The intelligent control method and car washing installation of car washing installation
CN108375911A (en) A kind of apparatus control method, device, storage medium and equipment
CN112626800B (en) Intelligent washing method and device and intelligent washing machine
CN111254682A (en) Ironing method, steam ironing device, system and storage medium
WO2024045601A1 (en) Ironing method and device for garment steamer, and garment steamer, storage medium and electronic device
WO2019036974A1 (en) Smart washing method, and washing machine
CN110894658B (en) Washing control method and device and clothes treatment device
US20230272569A1 (en) Clothing management apparatus and method for controlling thereof
WO2024021661A1 (en) Steaming control method and device of clothes steamer, and storage medium and electronic device
CN113622144B (en) Laundry control method, laundry control system, washing machine and computer readable storage medium
CN117671305A (en) Clothes care mode control method, system, storage medium and electronic device
CN115434103A (en) Washing machine control method and device and washing machine
CN113914062B (en) Clothes drying method and device, washing machine and storage medium
CN118087199A (en) Linkage method and device of towel rack and washing machine, electronic equipment and storage medium
CN114541082A (en) Clothes treatment equipment based on intelligent sound box and control method and control device thereof
CN118007357A (en) Method, device, equipment and medium for determining clothes washing scheme

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 23858642

Country of ref document: EP

Kind code of ref document: A1