CN106845556A - A kind of fabric defect detection method based on convolutional neural networks - Google Patents
A kind of fabric defect detection method based on convolutional neural networks Download PDFInfo
- Publication number
- CN106845556A CN106845556A CN201710072366.3A CN201710072366A CN106845556A CN 106845556 A CN106845556 A CN 106845556A CN 201710072366 A CN201710072366 A CN 201710072366A CN 106845556 A CN106845556 A CN 106845556A
- Authority
- CN
- China
- Prior art keywords
- image
- convolutional neural
- fault
- neural networks
- fabric
- Prior art date
- Legal status (The legal status 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 status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/30—Noise filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Probability & Statistics with Applications (AREA)
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
- Image Processing (AREA)
- Treatment Of Fiber Materials (AREA)
Abstract
The invention provides a kind of fabric defect detection method based on convolutional neural networks, textile image is obtained by image capture module, and send to image pre-processing module;Image pre-processing module is pre-processed to the textile image for being received, including image filtering treatment, compression of dynamic range treatment and histogram equalization processing;Convolutional neural networks model is set up, network inputs are pretreated textile image, and network is output as textile image with the presence or absence of fault and the situation of fault type;Pretreated textile image is input into set up convolutional neural networks model, convolutional neural networks automatically extract the different characteristic of textile image from multiple angles, and realization is identified and classifies to the fault in textile image.The method that the present invention is provided can be in each kind fabric of automatic detection fault, and fault to identifying classifies, and detection speed is fast, and testing result accuracy is high, and cost of implementation is low, is suitable to promote the use of on a large scale.
Description
Technical field
The present invention relates to image procossing and convolutional neural networks algorithm, specifically for detecting the fault in each kind fabric, and
Fabric defects to identifying is classified.
Background technology
China is weaving big country, and textile industry occupies critical role in social economy.Wherein, fabric quality is key,
The detection of fabric defects is even more the most important thing.At present, most of production line is also manually carrying out defect detection, artificial inspection
The speed of survey is slow, efficiency is low, it is easy to the subjective impact of examined personnel and missing inspection and false retrieval.
With continuing to develop for machine vision, image processing techniques and machine learning algorithm are also gradually applied to textile industry
In, the development of these technologies allows that defect detection realizes automation, so as to reach purpose quickly and efficiently.
To the research of Automatic Detection of Fabric Defects technology from the eighties in 20th century, more than 30 years, each national literature have been carried out
Person's consecutive publications substantial amounts of related article and achievement in research, achieve good achievement.At present, fabric defects is carried out automatically
The common methods of detection have image procossing, spectrum signature, wavelet transformation, Gabor filter, artificial neural network and genetic algorithms etc..
Although up to the present having occurred in that many good achievements in research, new method is continued to bring out out, steadily improve scientific research
Level, but more advanced automatic checkout system is after all a small number of.There is following defect in above-mentioned existing method:Differentiate fabric defects
Type it is limited in one's ability, the high cost of practical application, limitation is big.
Convolutional neural networks are a new research directions, fast-developing in machine learning field in recent years, are existed at present
Academia has been achieved in the multiclass application such as speech recognition, image recognition, natural language processing prominent by extensive concern
The progress of broken property.Can only be by reducing computing difficulty, convolutional neural networks algorithm with shallow structure relative to fully-connected network
Can be using the deep structure for possessing multilayer hidden layer, its advantage can be to complete complicated function approximation, be more suitable for a large amount of
The analysis of data and treatment work, because individual layer computing capability is limited, multitiered network can preferably obtain main structure letter
Breath.
The content of the invention
The technical problem to be solved in the present invention is how convolutional neural networks are applied to the detection and classification of fabric defects,
To improve the degree of accuracy and the efficiency of the resolution of fabric defects type.
In order to solve the above-mentioned technical problem, the technical scheme is that providing a kind of fabric based on convolutional neural networks
Defect detection method, it is characterised in that:The method is comprised the following steps:
Step 1:Textile image is obtained by image capture module, and is sent to image pre-processing module;
Step 2:Image pre-processing module is pre-processed to the textile image for being received, including image filtering treatment, pressure
The treatment of contracting dynamic range and histogram equalization processing;
Step 3:Convolutional neural networks model is set up, network inputs are pretreated textile image, and network is output as knitting
Object image is with the presence or absence of fault and the situation of fault type;
Step 4:The convolutional neural networks model that the pretreated textile image input step 3 of step 2 is set up, convolution
Neutral net automatically extracts the different characteristic of textile image from multiple angles, realization the fault in textile image is identified and
Classification.
Preferably, in the step 2, effectively filtered out in the case where artwork information is retained as far as possible using gaussian filtering and knitted
Noise in thing picture, is strengthened picture in the form of logarithmic transformation.
Preferably, in the step 3, convolutional neural networks model is included:Input layer, the fixed convolution of multilayered nonlinear
Layer, one layer of full articulamentum and output layer;Wherein, every layer of nonlinear fixed convolutional layer includes one layer of convolutional layer and one layer of pond again
Layer.
Preferably, in the step 3, the method for building up of convolutional neural networks model is as follows:
Step 3.1:Taken respectively by image capture module one group of intact textile image without fault, one group there is warp-wise
The textile image of fault, one group have the textile image of zonal fault, one group there is the textile image of domain type fault, one group of tool
There is the textile image of discrete type fault;
Step 3.2:The each group textile image in step 3.1 is pre-processed by image pre-processing module, including figure
As filtering process, compression of dynamic range treatment and histogram equalization processing;Pretreated each group textile image is divided respectively
Into two parts, used as training sample set, a part is used as test sample collection for a part;
Step 3.3:It is trained using training sample set pair convolutional neural networks model, if intact knitting without fault
Thing, then export the result of (0,0,0);If the fabric with warp fault, then the result of (0,0,1) is exported;If having latitude
To the fabric of fault, then the result of (0,1,0) is exported;If the fabric with domain type fault, then the knot of (1,0,0) is exported
Really;If the fabric with discrete type fault, then the result of (1,0,1) is exported;
Step 3.4:The convolutional neural networks model trained using test sample set pair step 3.3 is tested, and will be surveyed
Test result is contrasted with desired value, and such as error rate then returns to step 3.3 re -training not in the threshold range of setting, right
Network parameter is modified;Until error rate meets the threshold range of setting, obtain being adapted to the convolutional Neural of fabric defects detection
Network model.
Compared to existing technology, the fabric defect detection method based on convolutional neural networks that the present invention is provided has and has as follows
Beneficial effect:
The 1st, convolutional neural networks are used for the detection of fabric defects, convolutional neural networks can to describe the nonlinear dependence of complexity
System, but because its network it is openness greatly reduce its difficulty in computation, improve its computational efficiency so that more efficiently, it is accurate
Really it is fitted, improves the accuracy and efficiency of defect detection;
2nd, convolutional neural networks have the function of carrying out feature extraction automatically, and it can be from multiple angle extraction images not
Same feature, so as to solve the problems, such as how effectively to carry out feature extraction, realizing can not only be to the fault in textile image
It is identified, moreover it is possible to which the fault to identifying is classified;
3rd, image is pre-processed in taking pictures in real time, artwork information can be as far as possible being retained using gaussian filtering
In the case of filter noise in fabric picture well, compression of dynamic range and histogram equalization can be carried out to textile image
Enhancing, this is conducive to the identification and classification of fault;Processing routine is fixed, and is conducive to improving the speed of service;
4th, method realizes that conveniently, low cost is suitable to promote the use of on a large scale.
Brief description of the drawings
The fabric defect detection method flow chart based on convolutional neural networks that Fig. 1 is provided for the present embodiment;
Fig. 2 is convolutional neural networks structural representation.
Specific embodiment
With reference to specific embodiment, the present invention is expanded on further.It should be understood that these embodiments are merely to illustrate the present invention
Rather than limitation the scope of the present invention.In addition, it is to be understood that after the content for having read instruction of the present invention, people in the art
Member can make various changes or modifications to the present invention, and these equivalent form of values equally fall within the application appended claims and limited
Scope.
Fabric defects is detected and is classified, it is mainly included image capture module, image pre-processing module, convolution god
Through mixed-media network modules mixed-media.
Image capture module, mainly shooting fabric picture.
Image pre-processing module, mainly the fabric picture to shooting carry out image filtering treatment, the dynamic model of compression
Enclose and histogram equalization processing.
Convolutional neural networks model, is mainly identified and divided using convolutional neural networks to the fault in fabric picture
Class.In the network that the input of pretreated fabric picture is trained, fabric defects is detected.If intact fabric, then
Export the result of (0,0,0);If there is fault, also to exist fault classify, be broadly divided into warp fault (0,0,1),
Zonal fault (0,1,0), domain type fault (1,0,0) and discrete type fault (1,0,1).
With reference to Fig. 1, in the present embodiment, the fabric defect detection method flow based on convolutional neural networks is:
Step 1:Shoot 1100 different fabric picture samples at the scene using image capture module.
Step 2:Image filter is carried out to 1100 different fabric picture samples of shooting using image pre-processing module
Ripple treatment, with Gaussian filter by the noise filtering in image;Then dynamic range is compressed to image and histogram is equal
Weighing apparatusization treatment, is strengthened picture in the form of logarithmic transformation.
Step 3:Convolutional neural networks model is set up, the fault in fabric picture is entered using convolutional neural networks mainly
Row identification and classification.Wherein, training sample is above-mentioned textile image, and network inputs are textile image, automatically extract fabric figure
The characteristic vector of picture, network is output as textile image sample with the presence or absence of fault and the situation of fault type.Due to input sample
This numerical values recited differs, and gap is larger, it is necessary to carry out input data normalized.Secondly build network and carry out network
Initial work, sets 10 layers of nonlinear fixed convolutional layer, one layer of full articulamentum, wherein every layer of nonlinear fixed convolutional layer
Include one layer of convolutional layer and one layer of pond layer again.And to learning rate, training objective minimal error, maximum allowable train epochs and
Small lot data amount check is initialized.Then the training of convolutional neural networks is started, by substantial amounts of input and output training sample
In input network, the Weight Training in network is carried out, finally give the convolutional neural networks model with each new weights.Finally
Network test is carried out, test result is contrasted with desired value, to assess the training result of the network model, and to network
Parameter is modified.Finally, obtain being adapted to the convolutional neural networks detection model of fabric picture.
Image pre-processing module does not include gray processing process, that is, retain the coloured image of fabric, maintains as far as possible original
Pictorial information, is conducive to improving the degree of accuracy for detecting.
Image pre-processing module includes filtering, and the situation of artwork information can be as far as possible being retained using gaussian filtering
Under filter noise in fabric picture well, compression of dynamic range and histogram equalization can increase to textile image
By force, this is conducive to the identification and classification of fault.
Image pre-processing module is in taking pictures in real time, to image include image filtering, compression of dynamic range and directly
The operations such as side's figure equalization, its processing routine is fixed, and this is conducive to improving the speed of service.
The core algorithm of the inventive method is convolutional neural networks algorithm, when defect detection is carried out, uses convolutional Neural
Network model, it can describe the non-linear relation of complexity, and this is conducive to improving the accuracy of defect detection.Convolutional neural networks
With the function of carrying out feature extraction automatically, it can from the different characteristic of multiple angle extraction images, so as to solve how
The problem of feature extraction is effectively carried out, realization is identified and classifies to the fault in textile image.
Convolutional neural networks module is different from common BP neural network model, and general BP networks have one layer or two-layer
Hidden layer, structure is single, calculates simple;And the convolutional neural networks that this method is used can greatly improve its institute's energy descriptive system
Non-linear complexity, and because its network it is openness greatly reduce its difficulty in computation, improve its computational efficiency, from
And be more efficiently and accurately fitted.
As shown in Fig. 2 in the present embodiment, convolutional neural networks model is included:Input layer, 10 layers of nonlinear fixed convolution
Layer, one layer of full articulamentum and output layer, wherein, every layer of nonlinear fixed convolutional layer includes one layer of convolutional layer and one layer of pond again
Layer.
The step of convolutional neural networks algorithm, is as follows:
Step a, setting each initial parameter of neutral net, including the filtering of the nonlinear fixed convolutional layer number of plies, each convolutional layer
The number and its initial value of device, the window size of each pond layer, the full articulamentum number of plies and every layer of neuron number, build basic
Convolutional neural networks model;
Training objective minimal error, learning rate, maximum allowable train epochs and small lot in step b, setting model
Data amount check;
Step c, input layer is imported data to, the output of recursive calculation network;
Step d, calculating output error, and change weights.
Step e, repeat step a~step d, until error is in tolerance interval or beyond frequency of training limitation.
Convolutional neural networks model can effectively utilize the property modification weights of backpropagation, so as to preferably be intended
Close.
In the present embodiment, the method for building up for being adapted to the convolutional neural networks model of fabric defects detection is as follows:
Step 3.1:Taken respectively by image capture module one group of intact textile image without fault, one group there is warp-wise
The textile image of fault, one group have the textile image of zonal fault, one group there is the textile image of domain type fault, one group of tool
There is the textile image of discrete type fault;
Step 3.2:The each group textile image in step 3.1 is pre-processed by image pre-processing module, including figure
As filtering process, compression of dynamic range treatment and histogram equalization processing;Pretreated each group textile image is divided respectively
Into two parts, used as training sample set, a part is used as test sample collection for a part;
Step 3.3:It is trained using training sample set pair convolutional neural networks model, if intact knitting without fault
Thing, then export the result of (0,0,0);If the fabric with warp fault, then the result of (0,0,1) is exported;If having latitude
To the fabric of fault, then the result of (0,1,0) is exported;If the fabric with domain type fault, then the knot of (1,0,0) is exported
Really;If the fabric with discrete type fault, then the result of (1,0,1) is exported;
Step 3.4:The convolutional neural networks model trained using test sample set pair step 3.3 is tested, and will be surveyed
Test result is contrasted with desired value, and such as error rate then returns to step 3.3 re -training not in the threshold range of setting, right
Network parameter is modified;Until error rate meets the threshold range of setting, obtain being adapted to the convolutional Neural of fabric defects detection
Network model.
For convolutional neural networks model, it is also possible to first use and comprise only 0 or 1 one sample of numerical value output to convolutional Neural
Network of network model is trained and tests, and draws the network model that whether intact to fabric sample can be judged;Then,
Convolutional neural networks model is trained and tested with the sample exported containing three one-dimensional vectors of element, drawing can be right
The network model that the fault sample for identifying is classified.
Experiment shows that the fabric defects detection model based on convolutional neural networks that the present embodiment is set up being capable of automatic detection
Fault in each kind fabric, and fault to identifying classifies, and detection speed is fast, and testing result accuracy is high, fits
Promoted the use of on a large scale.
Claims (4)
1. a kind of fabric defect detection method based on convolutional neural networks, it is characterised in that:The method is comprised the following steps:
Step 1:Textile image is obtained by image capture module, and is sent to image pre-processing module;
Step 2:Image pre-processing module is pre-processed to the textile image for being received, including image filtering treatment, compression are moved
The treatment of state scope and histogram equalization processing;
Step 3:Convolutional neural networks model is set up, network inputs are pretreated textile image, and network is output as fabric figure
As the situation with the presence or absence of fault and fault type;
Step 4:The convolutional neural networks model that the pretreated textile image input step 3 of step 2 is set up, convolutional Neural
Network automatically extracts the different characteristic of textile image from multiple angles, and realization is identified and divides to the fault in textile image
Class.
2. a kind of fabric defect detection method based on convolutional neural networks as claimed in claim 1, it is characterised in that:It is described
In step 2, using noise of the gaussian filtering in fabric picture is effectively filtered out in the case of retaining artwork information as far as possible, use
The form of logarithmic transformation strengthens picture.
3. a kind of fabric defect detection method based on convolutional neural networks as claimed in claim 1, it is characterised in that:It is described
In step 3, convolutional neural networks model is included:Input layer, the fixed convolutional layer of multilayered nonlinear, one layer of full articulamentum and output
Layer;Wherein, every layer of nonlinear fixed convolutional layer includes one layer of convolutional layer and one layer of pond layer again.
4. a kind of fabric defect detection method based on convolutional neural networks as claimed in claim 1, it is characterised in that:It is described
In step 3, the method for building up of convolutional neural networks model is as follows:
Step 3.1:Taken respectively by image capture module one group of intact textile image without fault, one group there is warp fault
Textile image, one group have the textile image of zonal fault, one group have the textile image of domain type fault, one group have from
Dissipate the textile image of type fault;
Step 3.2:The each group textile image in step 3.1 is pre-processed by image pre-processing module, including image filter
Ripple treatment, compression of dynamic range treatment and histogram equalization processing;Pretreated each group textile image is respectively classified into two
Part, used as training sample set, a part is used as test sample collection for a part;
Step 3.3:It is trained using training sample set pair convolutional neural networks model, if the intact fabric without fault, then
Export the result of (0,0,0);If the fabric with warp fault, then the result of (0,0,1) is exported;If having zonal fault
Fabric, then export (0,1,0) result;If the fabric with domain type fault, then the result of (1,0,0) is exported;If
Fabric with discrete type fault, then export the result of (1,0,1);
Step 3.4:The convolutional neural networks model trained using test sample set pair step 3.3 is tested, and test is tied
Fruit is contrasted with desired value, and such as error rate then returns to step 3.3 re -training, to network not in the threshold range of setting
Parameter is modified;Until error rate meets the threshold range of setting, obtain being adapted to the convolutional neural networks of fabric defects detection
Model.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710072366.3A CN106845556A (en) | 2017-02-09 | 2017-02-09 | A kind of fabric defect detection method based on convolutional neural networks |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710072366.3A CN106845556A (en) | 2017-02-09 | 2017-02-09 | A kind of fabric defect detection method based on convolutional neural networks |
Publications (1)
Publication Number | Publication Date |
---|---|
CN106845556A true CN106845556A (en) | 2017-06-13 |
Family
ID=59122300
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710072366.3A Pending CN106845556A (en) | 2017-02-09 | 2017-02-09 | A kind of fabric defect detection method based on convolutional neural networks |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106845556A (en) |
Cited By (23)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107369155A (en) * | 2017-07-24 | 2017-11-21 | 广东工业大学 | A kind of cloth surface defect detection method and its system based on machine vision |
CN108154508A (en) * | 2018-01-09 | 2018-06-12 | 北京百度网讯科技有限公司 | Method, apparatus, storage medium and the terminal device of product defects detection positioning |
CN108242054A (en) * | 2018-01-09 | 2018-07-03 | 北京百度网讯科技有限公司 | A kind of steel plate defect detection method, device, equipment and server |
CN108389194A (en) * | 2018-02-24 | 2018-08-10 | 广州大久生物科技有限公司 | The image-recognizing method and device of denim water washing effect evaluation based on artificial intelligence |
CN108520212A (en) * | 2018-03-27 | 2018-09-11 | 东华大学 | Method for traffic sign detection based on improved convolutional neural networks |
CN109035248A (en) * | 2018-09-05 | 2018-12-18 | 深圳灵图慧视科技有限公司 | Defect detection method, apparatus, terminal device, server and storage medium |
CN109166118A (en) * | 2018-09-05 | 2019-01-08 | 深圳灵图慧视科技有限公司 | Fabric surface attribute detection method, device and computer equipment |
CN109187579A (en) * | 2018-09-05 | 2019-01-11 | 深圳灵图慧视科技有限公司 | Fabric defect detection method and device, computer equipment and computer-readable medium |
CN109242846A (en) * | 2018-09-05 | 2019-01-18 | 深圳灵图慧视科技有限公司 | Method, apparatus and equipment for fabric surface defects detection |
CN109272494A (en) * | 2018-08-31 | 2019-01-25 | 龙山县惹巴妹手工织品有限公司 | A kind of toy watch leather fabric detection method |
CN109283906A (en) * | 2018-11-10 | 2019-01-29 | 国网电力科学研究院武汉南瑞有限责任公司 | A kind of monitoring system and method stacking process |
CN109325940A (en) * | 2018-09-05 | 2019-02-12 | 深圳灵图慧视科技有限公司 | Textile detecting method and device, computer equipment and computer-readable medium |
CN109472735A (en) * | 2018-10-30 | 2019-03-15 | 深圳灵图慧视科技有限公司 | Realize accelerator, method and the acceleration system of fabric defects detection neural network |
WO2020010638A1 (en) * | 2018-07-13 | 2020-01-16 | 华为技术有限公司 | Method and device for detecting defective pixel in image |
CN111047655A (en) * | 2020-01-10 | 2020-04-21 | 北京盛开互动科技有限公司 | High-definition camera cloth defect detection method based on convolutional neural network |
CN111161207A (en) * | 2019-11-14 | 2020-05-15 | 西安工程大学 | Integrated convolutional neural network fabric defect classification method |
CN111429415A (en) * | 2020-03-18 | 2020-07-17 | 东华大学 | Efficient model construction method for product surface defects based on network collaborative pruning |
CN111504680A (en) * | 2020-04-30 | 2020-08-07 | 东华大学 | Fault diagnosis method and system for polyester filament yarn production based on WSVM and DCAE |
CN111751379A (en) * | 2020-06-30 | 2020-10-09 | 浙江双元科技开发有限公司 | Visual detection method and system for textile machine |
CN111815620A (en) * | 2020-07-24 | 2020-10-23 | 中原工学院 | Fabric defect detection method based on convolution characteristic and low-rank representation |
CN112036541A (en) * | 2020-10-16 | 2020-12-04 | 西安工程大学 | Fabric defect detection method based on genetic algorithm optimization neural network |
CN112215791A (en) * | 2019-07-12 | 2021-01-12 | 宝洁公司 | System and method for providing textile information and visualizing the same |
CN112465810A (en) * | 2020-12-15 | 2021-03-09 | 华南农业大学 | Method for detecting and classifying defects of textiles |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101598677A (en) * | 2009-07-03 | 2009-12-09 | 天津工业大学 | Online three-dimensional recognition method of fabric defects |
CN102879401A (en) * | 2012-09-07 | 2013-01-16 | 西安工程大学 | Method for automatically detecting and classifying textile flaws based on pattern recognition and image processing |
CN104751472A (en) * | 2015-04-10 | 2015-07-01 | 浙江工业大学 | Fabric defect detection method based on B-spline wavelets and deep neural network |
-
2017
- 2017-02-09 CN CN201710072366.3A patent/CN106845556A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101598677A (en) * | 2009-07-03 | 2009-12-09 | 天津工业大学 | Online three-dimensional recognition method of fabric defects |
CN102879401A (en) * | 2012-09-07 | 2013-01-16 | 西安工程大学 | Method for automatically detecting and classifying textile flaws based on pattern recognition and image processing |
CN104751472A (en) * | 2015-04-10 | 2015-07-01 | 浙江工业大学 | Fabric defect detection method based on B-spline wavelets and deep neural network |
Cited By (32)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107369155A (en) * | 2017-07-24 | 2017-11-21 | 广东工业大学 | A kind of cloth surface defect detection method and its system based on machine vision |
US10769774B2 (en) | 2018-01-09 | 2020-09-08 | Beijing Baidu Netcom Science And Technology Co., Ltd. | Method and device for detecting a defect in a steel plate, as well as apparatus and server therefor |
CN108154508A (en) * | 2018-01-09 | 2018-06-12 | 北京百度网讯科技有限公司 | Method, apparatus, storage medium and the terminal device of product defects detection positioning |
CN108242054A (en) * | 2018-01-09 | 2018-07-03 | 北京百度网讯科技有限公司 | A kind of steel plate defect detection method, device, equipment and server |
CN108389194A (en) * | 2018-02-24 | 2018-08-10 | 广州大久生物科技有限公司 | The image-recognizing method and device of denim water washing effect evaluation based on artificial intelligence |
CN108389194B (en) * | 2018-02-24 | 2020-11-17 | 广州大久生物科技有限公司 | Image recognition method and device based on artificial intelligence jean washing effect evaluation |
CN108520212A (en) * | 2018-03-27 | 2018-09-11 | 东华大学 | Method for traffic sign detection based on improved convolutional neural networks |
WO2020010638A1 (en) * | 2018-07-13 | 2020-01-16 | 华为技术有限公司 | Method and device for detecting defective pixel in image |
CN109272494B (en) * | 2018-08-31 | 2020-07-10 | 龙山县惹巴妹手工织品有限公司 | Method for detecting toy epidermis fabric |
CN109272494A (en) * | 2018-08-31 | 2019-01-25 | 龙山县惹巴妹手工织品有限公司 | A kind of toy watch leather fabric detection method |
CN109242846A (en) * | 2018-09-05 | 2019-01-18 | 深圳灵图慧视科技有限公司 | Method, apparatus and equipment for fabric surface defects detection |
CN109325940A (en) * | 2018-09-05 | 2019-02-12 | 深圳灵图慧视科技有限公司 | Textile detecting method and device, computer equipment and computer-readable medium |
CN109187579A (en) * | 2018-09-05 | 2019-01-11 | 深圳灵图慧视科技有限公司 | Fabric defect detection method and device, computer equipment and computer-readable medium |
WO2020048248A1 (en) * | 2018-09-05 | 2020-03-12 | 深圳灵图慧视科技有限公司 | Textile defect detection method and apparatus, and computer device and computer-readable medium |
CN109166118A (en) * | 2018-09-05 | 2019-01-08 | 深圳灵图慧视科技有限公司 | Fabric surface attribute detection method, device and computer equipment |
CN109035248A (en) * | 2018-09-05 | 2018-12-18 | 深圳灵图慧视科技有限公司 | Defect detection method, apparatus, terminal device, server and storage medium |
CN109472735A (en) * | 2018-10-30 | 2019-03-15 | 深圳灵图慧视科技有限公司 | Realize accelerator, method and the acceleration system of fabric defects detection neural network |
CN109283906A (en) * | 2018-11-10 | 2019-01-29 | 国网电力科学研究院武汉南瑞有限责任公司 | A kind of monitoring system and method stacking process |
CN112215791A (en) * | 2019-07-12 | 2021-01-12 | 宝洁公司 | System and method for providing textile information and visualizing the same |
WO2021008464A1 (en) * | 2019-07-12 | 2021-01-21 | The Procter & Gamble Company | System and method for providing textile information and visualizing same |
CN111161207A (en) * | 2019-11-14 | 2020-05-15 | 西安工程大学 | Integrated convolutional neural network fabric defect classification method |
CN111047655A (en) * | 2020-01-10 | 2020-04-21 | 北京盛开互动科技有限公司 | High-definition camera cloth defect detection method based on convolutional neural network |
CN111047655B (en) * | 2020-01-10 | 2024-05-14 | 北京盛开互动科技有限公司 | High-definition camera cloth defect detection method based on convolutional neural network |
CN111429415A (en) * | 2020-03-18 | 2020-07-17 | 东华大学 | Efficient model construction method for product surface defects based on network collaborative pruning |
CN111504680A (en) * | 2020-04-30 | 2020-08-07 | 东华大学 | Fault diagnosis method and system for polyester filament yarn production based on WSVM and DCAE |
CN111504680B (en) * | 2020-04-30 | 2021-03-26 | 东华大学 | Fault diagnosis method and system for polyester filament yarn production based on WSVM and DCAE |
CN111751379A (en) * | 2020-06-30 | 2020-10-09 | 浙江双元科技开发有限公司 | Visual detection method and system for textile machine |
CN111815620A (en) * | 2020-07-24 | 2020-10-23 | 中原工学院 | Fabric defect detection method based on convolution characteristic and low-rank representation |
CN111815620B (en) * | 2020-07-24 | 2021-05-04 | 中原工学院 | Fabric defect detection method based on convolution characteristic and low-rank representation |
CN112036541A (en) * | 2020-10-16 | 2020-12-04 | 西安工程大学 | Fabric defect detection method based on genetic algorithm optimization neural network |
CN112036541B (en) * | 2020-10-16 | 2023-11-17 | 西安工程大学 | Fabric defect detection method based on genetic algorithm optimization neural network |
CN112465810A (en) * | 2020-12-15 | 2021-03-09 | 华南农业大学 | Method for detecting and classifying defects of textiles |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106845556A (en) | A kind of fabric defect detection method based on convolutional neural networks | |
CN105303193B (en) | A kind of passenger number statistical system based on single-frame images processing | |
CN110717481B (en) | Method for realizing face detection by using cascaded convolutional neural network | |
CN107016405B (en) | A kind of pest image classification method based on classification prediction convolutional neural networks | |
CN107563999A (en) | A kind of chip defect recognition methods based on convolutional neural networks | |
CN106530288A (en) | Fabric defect detection method based on deep learning algorithm | |
CN109598287A (en) | The apparent flaws detection method that confrontation network sample generates is generated based on depth convolution | |
CN115205274A (en) | Cloth flaw detection method based on lightweight cascade network | |
CN105654066A (en) | Vehicle identification method and device | |
CN108416774A (en) | A kind of fabric types recognition methods based on fine granularity neural network | |
CN109359661A (en) | A kind of Sentinel-1 radar image classification method based on convolutional neural networks | |
CN104751186A (en) | Iris image quality classification method based on BP (back propagation) network and wavelet transformation | |
CN113506239B (en) | Strip steel surface defect detection method based on cross-stage local network | |
CN108764330A (en) | SAR image sorting technique based on super-pixel segmentation and convolution deconvolution network | |
CN115953666B (en) | Substation site progress identification method based on improved Mask-RCNN | |
CN111612759B (en) | Printed matter defect identification method based on deep convolution generation type countermeasure network | |
CN114863189B (en) | Intelligent image identification method based on big data | |
CN108960413A (en) | A kind of depth convolutional neural networks method applied to screw surface defects detection | |
Sakthimohan et al. | Detection and Recognition of Face Using Deep Learning | |
Jianli et al. | Identification of fabric defects based on discrete wavelet transform and back-propagation neural network | |
Guan et al. | Defect detection and classification for plain woven fabric based on deep learning | |
CN106940904A (en) | Attendance checking system based on recognition of face and speech recognition | |
CN110321936A (en) | A method of realizing that picture two is classified based on VGG16 and SVM | |
CN113379685A (en) | PCB defect detection method and device based on dual-channel feature comparison model | |
CN117197554A (en) | Transformer oil leakage real-time detection method and system |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170613 |
|
RJ01 | Rejection of invention patent application after publication |