CN116993733B - Earphone sleeve appearance quality detection method and system - Google Patents
Earphone sleeve appearance quality detection method and system Download PDFInfo
- Publication number
- CN116993733B CN116993733B CN202311253657.4A CN202311253657A CN116993733B CN 116993733 B CN116993733 B CN 116993733B CN 202311253657 A CN202311253657 A CN 202311253657A CN 116993733 B CN116993733 B CN 116993733B
- Authority
- CN
- China
- Prior art keywords
- target
- flaw
- level
- apparent
- detection surface
- 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.)
- Active
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 263
- 238000000034 method Methods 0.000 claims abstract description 34
- 238000003709 image segmentation Methods 0.000 claims abstract description 32
- 238000011156 evaluation Methods 0.000 claims description 60
- 230000000007 visual effect Effects 0.000 claims description 33
- 230000011218 segmentation Effects 0.000 claims description 26
- 238000013461 design Methods 0.000 claims description 25
- 238000012797 qualification Methods 0.000 claims description 25
- 238000012549 training Methods 0.000 claims description 22
- 238000004364 calculation method Methods 0.000 claims description 15
- 230000009467 reduction Effects 0.000 claims description 15
- 239000006185 dispersion Substances 0.000 claims description 9
- 230000002950 deficient Effects 0.000 claims description 7
- 238000013527 convolutional neural network Methods 0.000 claims description 4
- 230000003993 interaction Effects 0.000 claims description 4
- 238000013507 mapping Methods 0.000 claims description 4
- 238000009825 accumulation Methods 0.000 claims description 3
- 230000007547 defect Effects 0.000 claims description 2
- 238000007689 inspection Methods 0.000 claims 8
- 230000000694 effects Effects 0.000 abstract description 5
- 238000012986 modification Methods 0.000 description 4
- 230000004048 modification Effects 0.000 description 4
- 238000012545 processing Methods 0.000 description 4
- 230000008569 process Effects 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 230000010354 integration Effects 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 238000012216 screening Methods 0.000 description 2
- 239000013598 vector Substances 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 1
- 239000003086 colorant Substances 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000003708 edge detection Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000012372 quality testing Methods 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
- G06T7/0008—Industrial image inspection checking presence/absence
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D 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
- G06V10/10—Image acquisition
- G06V10/16—Image acquisition using multiple overlapping images; Image stitching
-
- 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
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
-
- 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
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
-
- 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
- G06V10/762—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using clustering, e.g. of similar faces in social networks
-
- 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
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
- G06V10/774—Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
-
- 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
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Software Systems (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Computing Systems (AREA)
- Databases & Information Systems (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Computer Graphics (AREA)
- Geometry (AREA)
- Quality & Reliability (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a method and a system for detecting the appearance quality of an earphone sleeve, and relates to the technical field of intelligent detection, wherein the method comprises the following steps: the method comprises the steps of constructing a target contour three-dimensional model, carrying out detection surface identification division to obtain a target multi-level detection surface, acquiring a target apparent image set through multi-dimensional image acquisition, carrying out image segmentation and clustering treatment on K target apparent images according to detection surface level identification to obtain an N-level apparent image set, synchronously inputting an appearance flaw identification sub-network to obtain N appearance flaw identification mark sets, and integrating to obtain a target appearance quality detection result. The invention solves the technical problems of the prior art that the quality detection stability and reliability are insufficient and the detection efficiency is low due to subjectivity of the detection of the quality of the headphone sleeve by using manpower, and achieves the technical effects of improving the stability, reliability and detection efficiency of the quality detection of the headphone sleeve.
Description
Technical Field
The invention relates to the technical field of intelligent detection, in particular to a method and a system for detecting the appearance quality of an earphone sleeve.
Background
The appearance flaw of the headset sleeve not only can cause bad product quality impression, but also can influence the product performance, so flaw detection is an important link of headset sleeve quality detection. However, at present, a manual mode is mostly adopted, and the quality of the headset sleeve is detected according to a preset detection evaluation rule, so that the stability and the reliability of quality detection are insufficient due to subjectivity of manual experience, and the detection efficiency is low.
Disclosure of Invention
The application provides a headset sleeve appearance quality detection method and system for use in solving among the prior art and carrying out the detection of headset sleeve quality by the manual work and have subjectivity, lead to quality detection stability and reliability not enough, and the low technical problem of detection efficiency.
In a first aspect of the present application, there is provided a method for detecting the quality of the external shape of an earphone sleeve, the method comprising: the method comprises the steps of obtaining target design information and target association information of a target earphone sleeve in an interaction mode; constructing a target contour three-dimensional model, and dividing detection surface identifiers of the target contour three-dimensional model according to the target association information to obtain a target multi-level detection surface, wherein the target contour three-dimensional model is constructed by taking the target design information as a reference, and the target multi-level detection surface comprises N levels of detection surfaces, wherein N is a positive integer; the target multi-stage detection surface projection is adopted to map the target earphone sleeve, so that a detection surface projection earphone sleeve is obtained; obtaining a target apparent image set, wherein the target apparent image set is obtained by carrying out multi-dimensional image acquisition on the detection surface projection earphone sleeve, the target apparent image set comprises K target apparent images, each target apparent image has a detection surface grade mark, and K is a positive integer; image segmentation of the target apparent image set is carried out according to the detection surface grade mark, and H apparent image segmentation results are obtained, wherein H is a positive integer greater than K; obtaining an N-level apparent image set, wherein the N-level apparent image set is obtained by clustering the H Zhang Biaoguan image segmentation result through the N-level detection surface; correspondingly synchronizing the N-level apparent image sets to an N-level apparent flaw identification sub-module of an appearance flaw identification sub-network to obtain N appearance flaw identification mark sets; and integrating the N appearance flaw identification mark sets to obtain a target appearance quality detection result.
In a second aspect of the present application, there is provided a headset sleeve appearance quality detection system, the system comprising: the target information acquisition module is used for interactively acquiring target design information and target association information of the target earphone sleeve; the target multi-level detection surface obtaining module is used for constructing a target contour three-dimensional model, and dividing detection surface identifiers of the target contour three-dimensional model according to the target association information to obtain a target multi-level detection surface, wherein the target contour three-dimensional model is constructed by taking the target design information as a reference, the target multi-level detection surface comprises N levels of detection surfaces, and N is a positive integer; the detection surface projection earphone sleeve acquisition module is used for mapping the target earphone sleeve by adopting the target multi-stage detection surface projection to obtain a detection surface projection earphone sleeve; the target apparent image set obtaining module is used for obtaining a target apparent image set, wherein the target apparent image set is obtained by carrying out multi-dimensional image acquisition on the detection surface projection earphone sleeve, the target apparent image set comprises K target apparent images, each target apparent image has a detection surface grade mark, and K is a positive integer; the apparent image segmentation module is used for carrying out image segmentation on the target apparent image set according to the detection surface grade identification to obtain H apparent image segmentation results, wherein H is a positive integer greater than K; the apparent image set obtaining module is used for obtaining an N-level apparent image set, wherein the N-level apparent image set is obtained by clustering the H Zhang Biaoguan image segmentation result through the N-level detection; the flaw identification mark set obtaining module is used for correspondingly synchronizing the N-level apparent image sets to an N-level apparent flaw identification sub-module of the appearance flaw identification sub-network to obtain N appearance flaw identification mark sets; the appearance quality detection result obtaining module is used for integrating the N appearance flaw identification mark sets to obtain a target appearance quality detection result.
One or more technical solutions provided in the present application have at least the following technical effects or advantages:
the utility model provides a headset appearance quality detection method relates to intelligent detection technical field, through constructing the three-dimensional model of target profile, carry out the detection face sign and divide, obtain the multistage detection face of target, through multidimensional image acquisition, obtain the target apparent image collection, including K target apparent image, carry out image segmentation and cluster processing to it according to detection face level sign, obtain N level apparent image collection, input the appearance flaw identification sub-network with it, obtain N appearance flaw identification mark collection, obtain target appearance quality testing result after the integration, the detection that uses the manpower to carry out headset appearance quality according to preset's detection evaluation rule in the prior art has subjectivity, result in quality detection stability and reliability not enough, and the low technical problem of detection efficiency, the stability of improvement headset appearance quality detection has been realized, reliability and detection efficiency's technical effect.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly described below, and it is apparent that the drawings in the following description are only some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a schematic flow chart of a method for detecting the appearance quality of an earphone sleeve according to an embodiment of the present application;
fig. 2 is a schematic flow chart of obtaining a target multi-stage detection surface in the earphone sleeve shape quality detection method according to the embodiment of the present application;
fig. 3 is a schematic flow chart of a method for detecting the quality of the appearance of an earphone sleeve according to an embodiment of the present application to obtain a target appearance quality detection result;
fig. 4 is a schematic structural diagram of an earphone sleeve appearance quality detection system according to an embodiment of the present application.
Reference numerals illustrate: the system comprises a target information acquisition module 11, a target multi-level detection surface acquisition module 12, a detection surface projection earphone sleeve acquisition module 13, a target apparent image set acquisition module 14, an apparent image segmentation module 15, an apparent image set acquisition module 16, a flaw identification mark set acquisition module 17 and an appearance quality detection result acquisition module 18.
Detailed Description
The application provides an earphone sleeve appearance quality detection method for use in solving among the prior art and carrying out the detection of earphone sleeve quality of headphone manually and have subjectivity, lead to quality detection stability and reliability not enough, and the low technical problem of detection efficiency.
The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. It will be apparent that the described embodiments are only some, but not all, of the embodiments of the present application. All other embodiments, which can be made by one of ordinary skill in the art without undue burden from the present disclosure, are within the scope of the present application based on the embodiments herein.
It should be noted that the terms "first," "second," and the like in the description and claims of the present application and the above figures are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged where appropriate such that embodiments of the present application described herein may be implemented in sequences other than those illustrated or otherwise described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or server that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed but may include other steps or modules not expressly listed or inherent to such process, method, article, or apparatus.
Example 1
As shown in fig. 1, the present application provides a method for detecting the appearance quality of an earphone sleeve, which includes:
s10: the method comprises the steps of obtaining target design information and target association information of a target earphone sleeve in an interaction mode;
specifically, target design information and target associated information of a target earphone sleeve are extracted interactively, wherein the interaction is data transmission, the target design information refers to product design information of the target earphone sleeve and comprises design parameters such as materials, outline shapes, sizes, appearance colors and the like of the target earphone sleeve, and the target associated information refers to relevant use information such as how the target earphone sleeve is matched with an earphone body for use, how the target earphone sleeve is mounted and attached and the like, and can be used for providing basic information for quality detection of the target earphone sleeve.
S20: constructing a target contour three-dimensional model, and dividing detection surface identifiers of the target contour three-dimensional model according to the target association information to obtain a target multi-level detection surface, wherein the target contour three-dimensional model is constructed by taking the target design information as a reference, and the target multi-level detection surface comprises N levels of detection surfaces, wherein N is a positive integer;
optionally, referring to the size, the material, and the like of the earphone sleeve in the target design information, constructing a three-dimensional virtual model of the target earphone sleeve by using virtual modeling software, extracting a model contour of the three-dimensional virtual model to obtain a target contour three-dimensional model, referring to the target association information, performing detection surface identification division on the target contour three-dimensional model, that is, dividing a quality detection surface of the earphone sleeve according to information such as the use function of the target earphone sleeve, and performing detection surface grade identification to obtain a target multi-stage detection surface, for example, taking an inner surface of the earphone sleeve, which is in contact with the human ear, as a type of quality detection surface, because the earphone sleeve is in direct contact with the human ear, the corresponding quality detection grade identification is higher, and the target multi-stage detection surface comprises N stages of detection surfaces which can be used as references for detecting the appearance quality of the target earphone sleeve.
Further, as shown in fig. 2, step S20 in the embodiment of the present application further includes:
s21: constructing an earphone sleeve reduction model, wherein the earphone sleeve reduction model is constructed and generated according to the target design information;
s22: extracting the contour of the earphone sleeve reduction model to obtain the target contour three-dimensional model;
s23: performing use fitting of the target earphone sleeve according to the target associated information to obtain a target detection surface division result, wherein the target detection surface division result comprises a first-stage visual detection surface, a second-stage contact detection surface and a third-stage hidden detection surface;
s24: and carrying out detection surface identification division of the target contour three-dimensional model by adopting the target detection surface division result to obtain a target multi-stage detection surface.
It should be understood that, according to the target design information, the design parameters of the target earphone sleeve are obtained, the earphone sleeve reduction model is constructed through virtual modeling software by using the design parameters, the contour extraction is carried out on the earphone sleeve reduction model, the contour of the object is identified and extracted by utilizing geometric calculation and mathematical algorithms such as edge detection, curve fitting, plane extraction and the like, the edge line or boundary curve of the object is found in the three-dimensional space of the earphone sleeve reduction model, and the edge line or boundary curve is further abstracted into the contour curve, so that the target contour three-dimensional model is obtained.
Further, performing use fitting of the target earphone sleeve according to the target related information, namely acquiring a use method of the target earphone sleeve, distinguishing use functions of each outer surface of the target earphone sleeve according to the use method, for example dividing the outermost layer of the earphone sleeve into a first-level visual detection surface, dividing the inner surface of the earphone sleeve, which is contacted with the human ear, into a second-level contact detection surface, dividing the inner surface of the earphone sleeve, which is in nested contact with the earphone body, into a third-level hidden detection surface, taking the first-level visual detection surface, the second-level contact detection surface and the third-level hidden detection surface as target detection surface division results, and performing corresponding detection surface identification division of the target contour three-dimensional model by adopting the target detection surface division results to obtain a target multi-level detection surface.
S30: the target multi-stage detection surface projection is adopted to map the target earphone sleeve, so that a detection surface projection earphone sleeve is obtained;
the target multi-level detection surfaces are adopted for projection mapping of the detection surfaces of the target earphone sleeve respectively to obtain projection earphone sleeves of the detection surfaces, namely the target multi-level detection surfaces are matched on the surfaces of the target earphone sleeve and positioned on all detection surfaces of the real object earphone sleeve, and the corresponding relation between the real object earphone sleeve and the target multi-level detection surfaces is determined so as to ensure accurate capturing of real appearance images of the real object earphone sleeve on all detection surfaces.
S40: obtaining a target apparent image set, wherein the target apparent image set is obtained by carrying out multi-dimensional image acquisition on the detection surface projection earphone sleeve, the target apparent image set comprises K target apparent images, each target apparent image has a detection surface grade mark, and K is a positive integer;
it should be understood that the image acquisition device, such as an ultra-fine camera, a thermal infrared imager and the like, is used for carrying out multi-dimensional image acquisition on the detection surface projection earphone sleeve, and the ultra-fine camera is used for acquiring appearance images of the target earphone sleeve from a plurality of shooting angles and a plurality of positions to obtain K target apparent images, each target apparent image has at least one detection surface grade identifier, and the target apparent image set is formed by the K target apparent images, so that the appearance quality condition of the target earphone sleeve on each detection surface can be reflected.
S50: image segmentation of the target apparent image set is carried out according to the detection surface grade mark, and H apparent image segmentation results are obtained, wherein H is a positive integer greater than K;
optionally, identifying a detection surface grade identifier of each target apparent image in the target apparent image set, and dividing the target apparent image according to the detection surface grade identifier, namely dividing the apparent images of different detection surfaces contained in one picture into different pictures, so that each picture has only one detection surface identifier, and the number of pictures is increased, therefore, after image division is performed on K target apparent images, H Zhang Biaoguan image division images with the number greater than K can be obtained, namely H apparent image division results.
S60: obtaining an N-level apparent image set, wherein the N-level apparent image set is obtained by clustering the H Zhang Biaoguan image segmentation result through the N-level detection surface;
in one possible embodiment of the present application, referring to the dividing standard of the N-level detection surface, the H Zhang Biaoguan image segmentation result is clustered, that is, the H Zhang Biaoguan image segmentation result is classified into three different clusters according to the first-level visual detection surface, the second-level contact detection surface and the third-level hidden detection surface, so as to obtain a first-level visual detection surface image set, a second-level contact detection surface image set and a third-level hidden detection surface image set, which are used as an N-level apparent image set as defect identification basic data of the target earphone sleeve.
S70: correspondingly synchronizing the N-level apparent image sets to an N-level apparent flaw identification sub-module of an appearance flaw identification sub-network to obtain N appearance flaw identification mark sets;
the apparent images in the N-level apparent image set are respectively input into N-level apparent flaw identification sub-modules corresponding to an appearance flaw identification sub-network, and appearance flaw identification is performed to obtain N appearance flaw identification mark sets, wherein the appearance flaw identification sub-network is a module for performing earphone sleeve appearance flaw identification, the appearance flaw identification mark sets can be obtained through training by using sample flaw data and a neural network calculation model, and the N appearance flaw identification mark sets can be integrated into an appearance quality detection result of a target earphone sleeve.
Further, step S70 in the embodiment of the present application further includes:
s71: acquiring a first segmentation training image set, wherein the first segmentation training image set corresponds to the primary visual detection surface;
s72: performing flaw feature identification marking on the first segmentation training image set to obtain a first flaw identification feature set;
s73: constructing a decoder unit and an encoder unit based on the convolutional neural network;
s74: training the decoder unit and the encoder unit using the first segmented training image set and the first flaw identification feature set;
s75: constructing a first-order apparent flaw identification submodule, wherein the first-order apparent flaw identification submodule is constructed and obtained based on the trained decoder unit and the trained encoder unit;
s76: and by analogy, a secondary apparent flaw identification sub-module for carrying out apparent flaw identification of the secondary contact detection surface is obtained;
s77: and the like, a three-level apparent flaw identification sub-module for carrying out apparent flaw identification on the three-level hidden detection surface is obtained;
s78: the first-stage apparent flaw identification sub-module, the second-stage apparent flaw identification sub-module and the third-stage apparent flaw identification sub-module form the N-stage apparent flaw identification sub-module;
S79: and correspondingly synchronizing the N-level apparent image sets to the N-level apparent flaw identification submodule to obtain the N apparent flaw identification mark sets.
For the first-stage visual detection surface, a plurality of sample apparent segmentation images are collected from a historical flaw detection case of the earphone sleeve, the sample apparent segmentation images comprise a plurality of marked normal images and marked flaw images, the normal images and the marked flaw images are used as a first segmentation training image set, flaw feature identification and marking are carried out on the images in the first segmentation training image set, namely all flaw features such as bulges, scratches, dirt, cracks and the like in the images are extracted, and marking is carried out according to flaw types, so that a first flaw identification feature set is formed. Further, the decoder unit and the encoder unit are constructed based on the architecture of a convolutional neural network, which is a deep neural network with a convolutional structure, comprising a plurality of layers of perceptrons, which are commonly used for analyzing visual images. The encoder unit is used for extracting flaw features of the input image, obtaining feature vectors of the image, and the decoder unit is used for converting the feature vectors extracted by the encoder unit into flaw identification marks of the corresponding image.
Further, the first segmentation training image set and the first flaw identification feature set are adopted as training data, and training, verification and testing are performed on the decoder unit and the encoder unit until the output of the decoder unit and the encoder unit reach convergence and meet the preset accuracy requirement, so that the training of the decoder unit and the encoder unit is completed. And constructing a first-stage apparent flaw identification sub-module based on the trained decoder unit and the trained encoder unit, wherein the first-stage apparent flaw identification sub-module is used for identifying the apparent flaws of the first-stage visual detection surface.
Similarly, a secondary apparent flaw identification sub-module for carrying out apparent flaw identification of the secondary contact detection surface and a tertiary apparent flaw identification sub-module for carrying out apparent flaw identification of the tertiary hidden detection surface are constructed. And the N-level apparent flaw identification sub-module is formed by the first-level apparent flaw identification sub-module, the second-level apparent flaw identification sub-module and the third-level apparent flaw identification sub-module.
Further, the first-stage visual detection surface image set, the second-stage contact detection surface image set and the third-stage hidden detection surface image set in the N-stage apparent flaw identification submodule are respectively and correspondingly synchronized to the N-stage apparent flaw identification submodule, and after flaw identification is carried out, the N apparent flaw identification mark sets are obtained.
S80: and integrating the N appearance flaw identification mark sets to obtain a target appearance quality detection result.
Further, as shown in fig. 3, step S80 in the embodiment of the present application further includes:
s81: the K target apparent images are provided with K apparent image acquisition position identifiers;
s82: estimating H segmentation image acquisition position identifiers of the H Zhang Biaoguan image segmentation result according to the K apparent image acquisition position identifiers;
s83: classifying the H segmentation image acquisition position identifiers according to the N-level apparent image set to obtain N-level segmentation image acquisition position identifiers;
s84: performing image stitching of the N appearance flaw identification mark sets correspondingly according to the N-level classification image acquisition position marks to obtain N-level appearance flaw stitched images;
s85: and the N-level appearance flaw spliced image forms the target appearance quality detection result.
Optionally, the K target apparent images are obtained based on a plurality of acquisition angles and a plurality of earphone sleeve positions, so that K apparent image acquisition position identifiers are respectively provided, the H Zhang Biaoguan image segmentation result is segmented from the K target apparent images, and therefore H segmentation image acquisition position identifiers of the H Zhang Biaoguan image segmentation result can be estimated according to the K apparent image acquisition position identifiers. Further, according to the classification standard of the N-level apparent image set, namely the detection surface level, the H segmentation image acquisition position identifiers are subjected to classification processing to obtain N-level segmentation image acquisition position identifiers.
Further, according to the N-level hierarchical image acquisition position identification, the images in the N-level visual flaw identification mark sets are spliced according to the corresponding positions to obtain N-level flaw spliced images, namely, a flaw spliced image of the primary visual detection surface, a flaw spliced image of the secondary contact detection surface and a flaw spliced image of the tertiary hidden detection surface, and the N-level visual flaw spliced images together serve as the target appearance quality detection result, so that appearance quality detection results of different parts of the target earphone sleeve can be reflected.
Further, step S85 in the embodiment of the present application further includes:
s85-1: pre-constructing a flaw type evaluation rule, wherein the flaw type evaluation rule comprises a primary flaw evaluation rule, a secondary flaw evaluation rule and a tertiary flaw evaluation rule;
s85-2: pre-constructing flaw evaluation weight distribution, wherein the flaw evaluation weight distribution comprises a primary flaw evaluation weight, a primary flaw evaluation weight and a tertiary flaw evaluation weight;
s85-3: traversing the N-level visual flaw spliced image based on the flaw type evaluation rule, and performing visual flaw identification mark scoring accumulation to obtain an N-level visual flaw scoring result;
S85-4: obtaining a target appearance quality detection score according to the N-level appearance flaw scoring result and the flaw evaluation weight distribution calculation;
s85-5: pre-constructing an earphone sleeve quality qualification threshold and an N-level quality qualification threshold;
s85-6: judging whether the target appearance quality detection score meets the earphone sleeve quality qualification threshold value and whether the N-level appearance flaw scoring result meets the N-level quality qualification threshold value;
s85-7: and if the target appearance quality detection score does not meet the earphone sleeve quality qualification threshold or the N-level appearance flaw scoring result does not meet the N-level quality qualification threshold, generating a defective product mark to identify the target earphone sleeve.
Specifically, according to different detection surface grades, different flaw type evaluation rules are pre-constructed, including a first-level flaw evaluation rule, a second-level flaw evaluation rule and a third-level flaw evaluation rule, and as the apparent quality requirements of the detection surfaces of different grades are different, the evaluation rules of the first-level flaw evaluation rule, the second-level flaw evaluation rule and the third-level flaw evaluation rule are also different, and for example, the second-level flaw evaluation rule corresponds to the second-level contact detection surface, so as to meet the comfort level of skin contact, the corresponding requirement of the second-level flaw evaluation rule is higher, the third-level flaw evaluation rule corresponds to the third-level hidden detection surface, the requirements of the appearance and the comfort level are relatively lower, and the corresponding evaluation rule requirements are also lower, for example, aiming at the same flaw, the second-level flaw evaluation rule is more buckled. Further, according to importance degrees of attractive appearance, comfort and the like of the product on the appearance quality of the earphone sleeve, flaw evaluation weight distribution is pre-built, namely, the primary flaw evaluation weight and the tertiary flaw evaluation weight corresponding to the primary flaw evaluation, the primary flaw evaluation and the tertiary flaw evaluation distribution are adopted.
Further, traversing the N-level visual flaw spliced image based on the flaw type evaluation rule, performing visual flaw identification marking, grading and accumulating according to flaw types to obtain N-level visual flaw grading results, namely primary flaw grading, primary flaw grading and tertiary flaw grading, and performing weighted calculation on the N-level visual flaw grading results in combination with the flaw evaluation weight distribution to obtain the target appearance quality detection grading.
Further, a headset sleeve quality qualification threshold and an N-level quality qualification threshold are respectively pre-constructed, namely, the overall quality detection score qualification standard value of a target headset sleeve and the quality detection score standard values of different detection surfaces are respectively pre-constructed, whether the target appearance quality detection score meets the headset sleeve quality qualification threshold and whether the N-level appearance flaw score result meets the N-level quality qualification threshold are judged, if the target appearance quality detection score does not meet the headset sleeve quality qualification threshold or the N-level appearance flaw score result does not meet the N-level quality qualification threshold at the same time, namely, when the overall quality of the target headset sleeve does not meet the requirement or the quality of each detection surface does not meet the requirement, a defective product mark is generated, and the target headset sleeve is marked as a defective product by using the defective product mark so as to achieve the purpose of screening out a defective product.
Further, step S85-3 of the embodiment of the present application further comprises:
s85-31: performing the same type flaw dispersion calculation on the N-level appearance flaw spliced image to obtain N groups of same type flaw dispersion indexes;
s85-32: obtaining N flaw discrete average values, wherein the N flaw discrete average values are generated by carrying out average value calculation on the N groups of flaw discrete indexes with the same type;
s85-33: and optimizing the N-level appearance flaw scoring result by adopting the N flaw discrete average values to obtain an N-level optimized flaw scoring result.
And aiming at the N-level visual flaw spliced image, respectively screening out the same type flaws, such as all stains, in the visual flaw spliced image of each level, and performing dispersion calculation aiming at the distribution positions of the same type flaws, wherein the dispersion degree of the same type flaws can be obtained by calculating the standard deviation of the distances among the same type flaws, so that N groups of flaw dispersion indexes of the same type are obtained. Further, average value calculation is performed on the N groups of flaw discrete indexes of the same type respectively to obtain N flaw discrete average values, the N flaw discrete average values are adopted to optimize the N-level appearance flaw grading result, that is, the flaw grading is corrected according to the distribution condition of flaws, so as to obtain an N-level optimized flaw grading result, and the larger the flaw discrete average value is, the more dispersed the flaw distribution is, the smaller the influence on the appearance quality is, and accordingly, the smaller the flaw discrete average value is, the more concentrated the flaw is, the larger the influence on the appearance quality is, and earphone sleeve appearance quality judgment is performed according to the N-level optimized flaw grading result, so that the accuracy of quality detection can be improved.
Further, the embodiment of the present application further includes step S90, where step S90 further includes:
s91: interactively obtaining a wiring characteristic area of the earphone sleeve;
s92: taking the earphone sleeve wiring characteristic region as an interested region to acquire local images of the K target apparent images, and obtaining a wiring characteristic image set;
s93: carrying out wiring error identification on the wiring characteristic image set to obtain a wiring error characteristic set, wherein each wiring error characteristic in the wiring error characteristic set has an error position mark;
s94: and adding the routing error feature set to the target appearance quality detection result.
It should be understood that the wiring characteristic area of the target earphone sleeve, that is, the setting and traversing area of the earphone wire, is identified, and the local image acquisition of the apparent image of the K targets is performed by taking the wiring characteristic area of the earphone sleeve as the region of interest, that is, the image of the wiring position of the earphone wire is acquired, so as to obtain the wiring characteristic image set. Further, according to the correct earphone wire routing standard, the routing characteristic image set is subjected to routing error recognition, namely, the characteristics of deviation of the earphone wire placement position, wrong threading position and the like are recognized, the routing error characteristic set is obtained, and the routing error characteristic set is added to the target appearance quality detection result, so that the purpose of recognizing the routing quality of the earphone sleeve is achieved.
In summary, the embodiments of the present application have at least the following technical effects:
according to the method, a three-dimensional model of a target contour is constructed to carry out detection surface identification division to obtain a target multi-stage detection surface, a target apparent image set is obtained through multi-dimensional image acquisition, the target multi-stage detection surface comprises K target apparent images, image segmentation and clustering processing are carried out on the target apparent images according to the detection surface grade identification to obtain an N-stage apparent image set, the N-stage apparent image set is input into an appearance flaw identification sub-network to obtain N appearance flaw identification mark sets, and a target appearance quality detection result is obtained after integration.
The technical effects of improving the stability, the reliability and the detection efficiency of the quality detection of the headset sleeve are achieved.
Example two
Based on the same inventive concept as the method for detecting the external shape quality of the earphone sleeve in the foregoing embodiments, as shown in fig. 4, the present application provides an earphone sleeve external shape quality detection system, and the system and method embodiments in the embodiments of the present application are based on the same inventive concept. Wherein the system comprises:
the target information acquisition module 11 is used for interactively acquiring target design information and target association information of the target earphone sleeve;
The target multi-level detection surface obtaining module 12 is configured to construct a target contour three-dimensional model, and perform detection surface identification division of the target contour three-dimensional model according to the target association information to obtain a target multi-level detection surface, where the target contour three-dimensional model is constructed by taking the target design information as a reference, and the target multi-level detection surface includes N-level detection surfaces, where N is a positive integer;
the detection surface projection earphone sleeve acquisition module 13 is used for mapping the target earphone sleeve by adopting the target multi-stage detection surface projection to obtain a detection surface projection earphone sleeve;
the target apparent image set obtaining module 14 is configured to obtain a target apparent image set, where the target apparent image set is obtained by performing multi-dimensional image acquisition on the detection surface projection earphone sleeve, and the target apparent image set includes K target apparent images, each target apparent image has a detection surface level identifier, and K is a positive integer;
the apparent image segmentation module 15 is used for carrying out image segmentation on the target apparent image set according to the detection surface grade identification to obtain H apparent image segmentation results, wherein H is a positive integer greater than K;
The apparent image set obtaining module 16 is configured to obtain an N-level apparent image set, where the N-level apparent image set is obtained by performing clustering on the H Zhang Biaoguan image segmentation result through the N-level detection;
the flaw identification mark set obtaining module 17, wherein the flaw identification mark set obtaining module 17 is used for correspondingly synchronizing the N-level apparent image set to an N-level apparent flaw identification sub-module of an appearance flaw identification sub-network to obtain N appearance flaw identification mark sets;
the appearance quality detection result obtaining module 18, the appearance quality detection result obtaining module 18 is configured to integrate the N appearance flaw identification mark sets to obtain a target appearance quality detection result.
Further, the target multi-level detection surface obtaining module 12 is further configured to perform the following steps:
constructing an earphone sleeve reduction model, wherein the earphone sleeve reduction model is constructed and generated according to the target design information;
extracting the contour of the earphone sleeve reduction model to obtain the target contour three-dimensional model;
performing use fitting of the target earphone sleeve according to the target associated information to obtain a target detection surface division result, wherein the target detection surface division result comprises a first-stage visual detection surface, a second-stage contact detection surface and a third-stage hidden detection surface;
And carrying out detection surface identification division of the target contour three-dimensional model by adopting the target detection surface division result to obtain a target multi-stage detection surface.
Further, the flaw identification mark set obtaining module 17 is further configured to perform the following steps:
acquiring a first segmentation training image set, wherein the first segmentation training image set corresponds to the primary visual detection surface;
performing flaw feature identification marking on the first segmentation training image set to obtain a first flaw identification feature set;
constructing a decoder unit and an encoder unit based on the convolutional neural network;
training the decoder unit and the encoder unit using the first segmented training image set and the first flaw identification feature set;
constructing a first-order apparent flaw identification submodule, wherein the first-order apparent flaw identification submodule is constructed and obtained based on the trained decoder unit and the trained encoder unit;
and by analogy, a secondary apparent flaw identification sub-module for carrying out apparent flaw identification of the secondary contact detection surface is obtained;
and the like, a three-level apparent flaw identification sub-module for carrying out apparent flaw identification on the three-level hidden detection surface is obtained;
The first-stage apparent flaw identification sub-module, the second-stage apparent flaw identification sub-module and the third-stage apparent flaw identification sub-module form the N-stage apparent flaw identification sub-module;
and correspondingly synchronizing the N-level apparent image sets to the N-level apparent flaw identification submodule to obtain the N apparent flaw identification mark sets.
Further, the appearance quality detection result obtaining module 18 is further configured to perform the following steps:
the K target apparent images are provided with K apparent image acquisition position identifiers;
estimating H segmentation image acquisition position identifiers of the H Zhang Biaoguan image segmentation result according to the K apparent image acquisition position identifiers;
classifying the H segmentation image acquisition position identifiers according to the N-level apparent image set to obtain N-level segmentation image acquisition position identifiers;
performing image stitching of the N appearance flaw identification mark sets correspondingly according to the N-level classification image acquisition position marks to obtain N-level appearance flaw stitched images;
and the N-level appearance flaw spliced image forms the target appearance quality detection result.
Further, the appearance quality detection result obtaining module 18 is further configured to perform the following steps:
Pre-constructing a flaw type evaluation rule, wherein the flaw type evaluation rule comprises a primary flaw evaluation rule, a secondary flaw evaluation rule and a tertiary flaw evaluation rule;
pre-constructing flaw evaluation weight distribution, wherein the flaw evaluation weight distribution comprises a primary flaw evaluation weight, a primary flaw evaluation weight and a tertiary flaw evaluation weight;
traversing the N-level visual flaw spliced image based on the flaw type evaluation rule, and performing visual flaw identification mark scoring accumulation to obtain an N-level visual flaw scoring result;
obtaining a target appearance quality detection score according to the N-level appearance flaw scoring result and the flaw evaluation weight distribution calculation;
pre-constructing an earphone sleeve quality qualification threshold and an N-level quality qualification threshold;
judging whether the target appearance quality detection score meets the earphone sleeve quality qualification threshold value and whether the N-level appearance flaw scoring result meets the N-level quality qualification threshold value;
and if the target appearance quality detection score does not meet the earphone sleeve quality qualification threshold or the N-level appearance flaw scoring result does not meet the N-level quality qualification threshold, generating a defective product mark to identify the target earphone sleeve.
Further, the appearance quality detection result obtaining module 18 is further configured to perform the following steps:
performing the same type flaw dispersion calculation on the N-level appearance flaw spliced image to obtain N groups of same type flaw dispersion indexes;
obtaining N flaw discrete average values, wherein the N flaw discrete average values are generated by carrying out average value calculation on the N groups of flaw discrete indexes with the same type;
and optimizing the N-level appearance flaw scoring result by adopting the N flaw discrete average values to obtain an N-level optimized flaw scoring result.
Further, the system further comprises:
the wiring characteristic region acquisition module is used for interactively acquiring a wiring characteristic region of the earphone sleeve;
the wiring characteristic image set acquisition module is used for acquiring local images of the K target apparent images by taking the earphone sleeve wiring characteristic region as an interested region to obtain a wiring characteristic image set;
the wiring error feature set acquisition module is used for carrying out wiring error identification on the wiring feature image set to obtain a wiring error feature set, wherein each wiring error feature in the wiring error feature set is provided with an error position mark;
And the wiring error feature adding module is used for adding the wiring error feature set to the target appearance quality detection result.
It should be noted that the sequence of the embodiments of the present application is merely for description, and does not represent the advantages and disadvantages of the embodiments. And the foregoing description has been directed to specific embodiments of this specification. Other embodiments are within the scope of the following claims. In some cases, the actions or steps recited in the claims can be performed in a different order than in the embodiments and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
The foregoing description of the preferred embodiments of the present application is not intended to limit the invention to the particular embodiments of the present application, but to limit the scope of the invention to the particular embodiments of the present application.
The specification and drawings are merely exemplary of the application and are to be regarded as covering any and all modifications, variations, combinations, or equivalents that are within the scope of the application. It will be apparent to those skilled in the art that various modifications and variations can be made in the present application without departing from the scope of the application. Thus, if such modifications and variations of the present application fall within the scope of the present application and the equivalents thereof, the present application is intended to cover such modifications and variations.
Claims (7)
1. A method for detecting the appearance quality of an earphone sleeve, which is characterized by comprising the following steps:
the method comprises the steps of obtaining target design information and target association information of a target earphone sleeve in an interaction mode;
constructing a target contour three-dimensional model, and dividing detection surface identifiers of the target contour three-dimensional model according to the target association information to obtain a target multi-level detection surface, wherein the target contour three-dimensional model is constructed by taking the target design information as a reference, and the target multi-level detection surface comprises N levels of detection surfaces, wherein N is a positive integer;
the target multi-stage detection surface projection is adopted to map the target earphone sleeve, so that a detection surface projection earphone sleeve is obtained;
obtaining a target apparent image set, wherein the target apparent image set is obtained by carrying out multi-dimensional image acquisition on the detection surface projection earphone sleeve, the target apparent image set comprises K target apparent images, each target apparent image has a detection surface grade mark, and K is a positive integer;
image segmentation of the target apparent image set is carried out according to the detection surface grade mark, and H apparent image segmentation results are obtained, wherein H is a positive integer greater than K;
obtaining an N-level apparent image set, wherein the N-level apparent image set is obtained by clustering the H Zhang Biaoguan image segmentation result through the N-level detection surface;
Correspondingly synchronizing the N-level apparent image sets to an N-level apparent flaw identification sub-module of an appearance flaw identification sub-network to obtain N appearance flaw identification mark sets;
integrating the N appearance flaw identification mark sets to obtain a target appearance quality detection result;
the method comprises the steps of constructing a target contour three-dimensional model, dividing detection surface identifiers of the target contour three-dimensional model according to target association information to obtain a target multi-level detection surface, constructing the target contour three-dimensional model by taking target design information as a reference, wherein the target multi-level detection surface comprises N levels of detection surfaces, N is a positive integer, and further comprising:
constructing an earphone sleeve reduction model, wherein the earphone sleeve reduction model is constructed and generated according to the target design information;
extracting the contour of the earphone sleeve reduction model to obtain the target contour three-dimensional model;
performing use fitting of the target earphone sleeve according to the target associated information to obtain a target detection surface division result, wherein the target detection surface division result comprises a first-stage visual detection surface, a second-stage contact detection surface and a third-stage hidden detection surface, and the use fitting of the target earphone sleeve according to the target associated information refers to obtaining a use method of the target earphone sleeve and distinguishing use functions of each outer surface of the target earphone sleeve according to the use method;
And carrying out detection surface identification division of the target contour three-dimensional model by adopting the target detection surface division result to obtain a target multi-stage detection surface.
2. The quality inspection method of claim 1, wherein the synchronizing the N-level apparent image sets to the N-level apparent flaw identification submodule of the appearance flaw identification submodule to obtain N apparent flaw identification mark sets, further comprises:
acquiring a first segmentation training image set, wherein the first segmentation training image set corresponds to the primary visual detection surface;
performing flaw feature identification marking on the first segmentation training image set to obtain a first flaw identification feature set;
constructing a decoder unit and an encoder unit based on the convolutional neural network;
training the decoder unit and the encoder unit using the first segmented training image set and the first flaw identification feature set;
constructing a first-order apparent flaw identification submodule, wherein the first-order apparent flaw identification submodule is constructed and obtained based on the trained decoder unit and the trained encoder unit;
and by analogy, a secondary apparent flaw identification sub-module for carrying out apparent flaw identification of the secondary contact detection surface is obtained;
And the like, a three-level apparent flaw identification sub-module for carrying out apparent flaw identification on the three-level hidden detection surface is obtained;
the first-stage apparent flaw identification sub-module, the second-stage apparent flaw identification sub-module and the third-stage apparent flaw identification sub-module form the N-stage apparent flaw identification sub-module;
and correspondingly synchronizing the N-level apparent image sets to the N-level apparent flaw identification submodule to obtain the N apparent flaw identification mark sets.
3. The quality inspection method of claim 2, wherein integrating the N sets of appearance defect identification marks to obtain a target appearance quality inspection result further comprises:
the K target apparent images are provided with K apparent image acquisition position identifiers;
estimating H segmentation image acquisition position identifiers of the H Zhang Biaoguan image segmentation result according to the K apparent image acquisition position identifiers;
classifying the H segmentation image acquisition position identifiers according to the N-level apparent image set to obtain N-level segmentation image acquisition position identifiers;
performing image stitching of the N appearance flaw identification mark sets correspondingly according to the N-level classification image acquisition position marks to obtain N-level appearance flaw stitched images;
And the N-level appearance flaw spliced image forms the target appearance quality detection result.
4. A quality inspection method according to claim 3, further comprising:
pre-constructing a flaw type evaluation rule, wherein the flaw type evaluation rule comprises a primary flaw evaluation rule, a secondary flaw evaluation rule and a tertiary flaw evaluation rule;
pre-constructing flaw evaluation weight distribution, wherein the flaw evaluation weight distribution comprises a primary flaw evaluation weight, a primary flaw evaluation weight and a tertiary flaw evaluation weight;
traversing the N-level visual flaw spliced image based on the flaw type evaluation rule, and performing visual flaw identification mark scoring accumulation to obtain an N-level visual flaw scoring result;
obtaining a target appearance quality detection score according to the N-level appearance flaw scoring result and the flaw evaluation weight distribution calculation;
pre-constructing an earphone sleeve quality qualification threshold and an N-level quality qualification threshold;
judging whether the target appearance quality detection score meets the earphone sleeve quality qualification threshold value and whether the N-level appearance flaw scoring result meets the N-level quality qualification threshold value;
And if the target appearance quality detection score does not meet the earphone sleeve quality qualification threshold or the N-level appearance flaw scoring result does not meet the N-level quality qualification threshold, generating a defective product mark to identify the target earphone sleeve.
5. The quality inspection method according to claim 4, wherein the quality inspection method further comprises, before obtaining a target appearance quality inspection score from the N-level appearance flaw score result and the flaw evaluation weight assignment calculation:
performing the same type flaw dispersion calculation on the N-level appearance flaw spliced image to obtain N groups of same type flaw dispersion indexes;
obtaining N flaw discrete average values, wherein the N flaw discrete average values are generated by carrying out average value calculation on the N groups of flaw discrete indexes with the same type;
and optimizing the N-level appearance flaw scoring result by adopting the N flaw discrete average values to obtain an N-level optimized flaw scoring result.
6. The quality inspection method according to claim 1, further comprising:
interactively obtaining a wiring characteristic area of the earphone sleeve;
taking the earphone sleeve wiring characteristic region as an interested region to acquire local images of the K target apparent images, and obtaining a wiring characteristic image set;
Carrying out wiring error identification on the wiring characteristic image set to obtain a wiring error characteristic set, wherein each wiring error characteristic in the wiring error characteristic set has an error position mark;
and adding the routing error feature set to the target appearance quality detection result.
7. A headset case appearance quality detection system, the system comprising:
the target information acquisition module is used for interactively acquiring target design information and target association information of the target earphone sleeve;
the target multi-level detection surface obtaining module is used for constructing a target contour three-dimensional model, and dividing detection surface identifiers of the target contour three-dimensional model according to the target association information to obtain a target multi-level detection surface, wherein the target contour three-dimensional model is constructed by taking the target design information as a reference, the target multi-level detection surface comprises N levels of detection surfaces, and N is a positive integer;
the detection surface projection earphone sleeve acquisition module is used for mapping the target earphone sleeve by adopting the target multi-stage detection surface projection to obtain a detection surface projection earphone sleeve;
The target apparent image set obtaining module is used for obtaining a target apparent image set, wherein the target apparent image set is obtained by carrying out multi-dimensional image acquisition on the detection surface projection earphone sleeve, the target apparent image set comprises K target apparent images, each target apparent image has a detection surface grade mark, and K is a positive integer;
the apparent image segmentation module is used for carrying out image segmentation on the target apparent image set according to the detection surface grade identification to obtain H apparent image segmentation results, wherein H is a positive integer greater than K;
the apparent image set obtaining module is used for obtaining an N-level apparent image set, wherein the N-level apparent image set is obtained by clustering the H Zhang Biaoguan image segmentation result through the N-level detection;
the flaw identification mark set obtaining module is used for correspondingly synchronizing the N-level apparent image sets to an N-level apparent flaw identification sub-module of the appearance flaw identification sub-network to obtain N appearance flaw identification mark sets;
The appearance quality detection result obtaining module is used for integrating the N appearance flaw identification mark sets to obtain a target appearance quality detection result;
the target multi-stage detection surface obtaining module is further used for executing the following steps:
constructing an earphone sleeve reduction model, wherein the earphone sleeve reduction model is constructed and generated according to the target design information;
extracting the contour of the earphone sleeve reduction model to obtain the target contour three-dimensional model;
performing use fitting of the target earphone sleeve according to the target associated information to obtain a target detection surface division result, wherein the target detection surface division result comprises a first-stage visual detection surface, a second-stage contact detection surface and a third-stage hidden detection surface, and the use fitting of the target earphone sleeve according to the target associated information refers to obtaining a use method of the target earphone sleeve and distinguishing use functions of each outer surface of the target earphone sleeve according to the use method;
and carrying out detection surface identification division of the target contour three-dimensional model by adopting the target detection surface division result to obtain a target multi-stage detection surface.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202311253657.4A CN116993733B (en) | 2023-09-27 | 2023-09-27 | Earphone sleeve appearance quality detection method and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202311253657.4A CN116993733B (en) | 2023-09-27 | 2023-09-27 | Earphone sleeve appearance quality detection method and system |
Publications (2)
Publication Number | Publication Date |
---|---|
CN116993733A CN116993733A (en) | 2023-11-03 |
CN116993733B true CN116993733B (en) | 2024-01-09 |
Family
ID=88534204
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202311253657.4A Active CN116993733B (en) | 2023-09-27 | 2023-09-27 | Earphone sleeve appearance quality detection method and system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN116993733B (en) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN118025597A (en) * | 2024-04-01 | 2024-05-14 | 中钠储能新能源科技(甘肃)有限公司 | On-line detection method and system for lithium battery energy storage battery package |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2014025809A (en) * | 2012-07-26 | 2014-02-06 | Jfe Steel Corp | Method and apparatus for detecting scratch |
CN111812106A (en) * | 2020-09-15 | 2020-10-23 | 沈阳风驰软件股份有限公司 | Method and system for detecting glue overflow of appearance surface of wireless earphone |
CN115601355A (en) * | 2022-11-10 | 2023-01-13 | 四川启睿克科技有限公司(Cn) | Method and device for detecting and classifying product surface defects and storage medium |
CN116152209A (en) * | 2023-02-27 | 2023-05-23 | 中国地质大学(武汉) | Earphone cover defect detection method, device, equipment and storage medium |
CN116754484A (en) * | 2023-06-19 | 2023-09-15 | 江苏省特种设备安全监督检验研究院 | Nondestructive testing method for nonmetallic liner fiber winding container |
-
2023
- 2023-09-27 CN CN202311253657.4A patent/CN116993733B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2014025809A (en) * | 2012-07-26 | 2014-02-06 | Jfe Steel Corp | Method and apparatus for detecting scratch |
CN111812106A (en) * | 2020-09-15 | 2020-10-23 | 沈阳风驰软件股份有限公司 | Method and system for detecting glue overflow of appearance surface of wireless earphone |
CN115601355A (en) * | 2022-11-10 | 2023-01-13 | 四川启睿克科技有限公司(Cn) | Method and device for detecting and classifying product surface defects and storage medium |
CN116152209A (en) * | 2023-02-27 | 2023-05-23 | 中国地质大学(武汉) | Earphone cover defect detection method, device, equipment and storage medium |
CN116754484A (en) * | 2023-06-19 | 2023-09-15 | 江苏省特种设备安全监督检验研究院 | Nondestructive testing method for nonmetallic liner fiber winding container |
Non-Patent Citations (2)
Title |
---|
A Method of Surface Defect Detection of Bluetooth Headset Based on Machine Vision;Xin Lu et al.;2021 IEEE 24th International Conference on Computational Science and Engineering (CSE);第1-6页 * |
基于智能手机单目视觉三维重建的小型金属件测量方法;何文韬 等;软件;41(08);第5-9、20页 * |
Also Published As
Publication number | Publication date |
---|---|
CN116993733A (en) | 2023-11-03 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109829914B (en) | Method and device for detecting product defects | |
CN106599854B (en) | Automatic facial expression recognition method based on multi-feature fusion | |
CN114757900B (en) | Artificial intelligence-based textile defect type identification method | |
CN105740780B (en) | Method and device for detecting living human face | |
US20160196467A1 (en) | Three-Dimensional Face Recognition Device Based on Three Dimensional Point Cloud and Three-Dimensional Face Recognition Method Based on Three-Dimensional Point Cloud | |
CN116993733B (en) | Earphone sleeve appearance quality detection method and system | |
CN105975941A (en) | Multidirectional vehicle model detection recognition system based on deep learning | |
CN105574550A (en) | Vehicle identification method and device | |
CN108629319B (en) | Image detection method and system | |
CN112200045A (en) | Remote sensing image target detection model establishing method based on context enhancement and application | |
CN103714077A (en) | Method and device for retrieving objects and method and device for verifying retrieval | |
CN106557740B (en) | The recognition methods of oil depot target in a kind of remote sensing images | |
CN112529901B (en) | Crack identification method in complex environment | |
CN115272312B (en) | Plastic mobile phone shell defect detection method based on machine vision | |
CN111242899A (en) | Image-based flaw detection method and computer-readable storage medium | |
CN111758117A (en) | Inspection system, recognition system, and learning data generation device | |
CN114092478B (en) | Anomaly detection method | |
CN115294109A (en) | Real wood board production defect identification system based on artificial intelligence, and electronic equipment | |
CN109344758B (en) | Face recognition method based on improved local binary pattern | |
CN111178405A (en) | Similar object identification method fusing multiple neural networks | |
WO2022222036A1 (en) | Method and apparatus for determining parking space | |
CN113989196A (en) | Vision-based earphone silica gel gasket appearance defect detection method | |
CN115131539B (en) | Aluminum template automatic identification and classification system based on machine vision | |
CN111723628A (en) | Figure re-recognition method, figure re-recognition system and image screening method | |
CN108256578B (en) | Gray level image identification method, device, equipment and readable storage medium |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant |