CN116309493A - Method and system for detecting defects of textile products - Google Patents
Method and system for detecting defects of textile products Download PDFInfo
- Publication number
- CN116309493A CN116309493A CN202310298187.7A CN202310298187A CN116309493A CN 116309493 A CN116309493 A CN 116309493A CN 202310298187 A CN202310298187 A CN 202310298187A CN 116309493 A CN116309493 A CN 116309493A
- Authority
- CN
- China
- Prior art keywords
- defect
- textile product
- image
- textile
- product defect
- 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
- 230000007547 defect Effects 0.000 title claims abstract description 443
- 239000004753 textile Substances 0.000 title claims abstract description 263
- 238000000034 method Methods 0.000 title claims abstract description 45
- 238000000605 extraction Methods 0.000 claims abstract description 64
- 238000001514 detection method Methods 0.000 claims abstract description 45
- 238000007621 cluster analysis Methods 0.000 claims abstract description 34
- 230000002159 abnormal effect Effects 0.000 claims abstract description 10
- 230000009466 transformation Effects 0.000 claims description 25
- 238000012549 training Methods 0.000 claims description 23
- 238000012360 testing method Methods 0.000 claims description 14
- 238000013135 deep learning Methods 0.000 claims description 12
- 238000009826 distribution Methods 0.000 claims description 11
- 230000007613 environmental effect Effects 0.000 claims description 10
- 238000012545 processing Methods 0.000 claims description 10
- 238000004422 calculation algorithm Methods 0.000 claims description 9
- 238000007781 pre-processing Methods 0.000 claims description 9
- 238000012795 verification Methods 0.000 claims description 9
- 230000003321 amplification Effects 0.000 claims description 8
- 238000004519 manufacturing process Methods 0.000 claims description 8
- 238000003199 nucleic acid amplification method Methods 0.000 claims description 8
- 238000004458 analytical method Methods 0.000 claims description 7
- 238000006243 chemical reaction Methods 0.000 claims description 7
- 238000001914 filtration Methods 0.000 claims description 4
- 238000013507 mapping Methods 0.000 claims description 4
- 230000008485 antagonism Effects 0.000 claims description 3
- 238000005457 optimization Methods 0.000 claims description 2
- 230000000694 effects Effects 0.000 abstract description 15
- 238000013473 artificial intelligence Methods 0.000 abstract description 2
- 230000008569 process Effects 0.000 description 5
- 238000004364 calculation method Methods 0.000 description 3
- 230000002950 deficient Effects 0.000 description 3
- 230000002708 enhancing effect Effects 0.000 description 3
- 239000004744 fabric Substances 0.000 description 2
- 238000007689 inspection Methods 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 230000000007 visual effect Effects 0.000 description 2
- 230000015572 biosynthetic process Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 239000006185 dispersion Substances 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000010191 image analysis Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000009987 spinning Methods 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
- 230000001131 transforming effect Effects 0.000 description 1
Images
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
-
- 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
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
- G06T2207/30124—Fabrics; Textile; Paper
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computing Systems (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Databases & Information Systems (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Multimedia (AREA)
- Quality & Reliability (AREA)
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
- Treatment Of Fiber Materials (AREA)
Abstract
The application discloses a method and a system for detecting defects of textile products, which belong to the field of artificial intelligence, and the method comprises the following steps: acquiring a textile product defect image set through big data acquisition, constructing a textile product defect tag library, acquiring textile product defect tag set information, carrying out cluster analysis on the image defects, acquiring a textile product defect cluster analysis result, carrying out abnormal point feature extraction, acquiring a defect feature extraction value, carrying out region merging on the defect extraction region set based on the textile product defect cluster analysis result, acquiring a defect effective region set, determining a textile product defect feature set, and detecting and identifying textile product defect images to be processed by using a textile product defect identification model. Solves the technical problems of lack of efficient detection of textile product defects and low detection quality and efficiency in the prior art. The technical effects of intelligently identifying defects in textiles and improving identification efficiency and accuracy are achieved.
Description
Technical Field
The application relates to the field of artificial intelligence, in particular to a method and a system for detecting defects of textile products.
Background
China is a large country for textile production and export, and the textile industry develops for many years to form a complete industrial chain and a matched high processing level. Along with the continuous change of market demands, research on production and processing technology of the textile industry is promoted, and the method has very important significance for meeting challenges brought by the market.
Currently, textiles must undergo a wide variety of inspections and tests during production and prior to entering the market, with defect monitoring being the most prominent component. Often, the detection of fabric defects is mainly accomplished by manual visual off-line detection. The number of defects in the unit mass of fabric was counted by eye examination.
However, since the defect statistics result is greatly influenced by the subjective influence of the inspector through manual inspection, the false detection rate and the omission rate are relatively high. The defects in the textiles are not sufficiently detected, the error of the statistical result is larger, meanwhile, the efficiency is lower due to manual detection, and the detection speed is lower. There is a technical problem of lack of efficient detection of defects in textile products, and low detection quality and efficiency.
Disclosure of Invention
The purpose of the application is to provide a method and a system for detecting defects of textile products, which are used for solving the technical problems of lack of efficient detection of defects of textile products, low detection quality and low efficiency in the prior art.
In view of the foregoing, the present application provides a method and system for detecting defects in textile products.
In a first aspect, the present application provides a method for defect detection of textile products, wherein the method comprises: acquiring a textile product defect image set through big data acquisition; constructing a textile product defect label library, and classifying the defect labels of the defect images based on the textile product defect label library to obtain textile product defect label set information; performing cluster analysis on the image defects according to the textile product defect label set information to obtain a textile product defect cluster analysis result; performing abnormal point feature extraction on each defect image in the defect image set of the textile product to obtain a defect feature extraction value; acquiring a defect extraction area set of the defect feature extraction value, and carrying out area combination on the defect extraction area set based on a textile product defect cluster analysis result to obtain a defect effective area set; determining a defect feature set of the textile product according to the defect feature extraction value and the defect effective area set; performing model training on the textile product defect feature set by using a deep learning network structure to obtain a textile product defect recognition model; and detecting and identifying the defect image of the textile product to be treated based on the defect identification model of the textile product.
In another aspect, the present application also provides a system for defect detection in textile products, wherein the system comprises: the image acquisition module is used for acquiring a textile product defect image set through big data acquisition; the label information obtaining module is used for constructing a textile product defect label library, classifying the defect labels of the defect images based on the textile product defect label library, and obtaining textile product defect label set information; the cluster analysis module is used for carrying out cluster analysis on the image defects according to the textile product defect label set information to obtain a textile product defect cluster analysis result; the feature extraction module is used for carrying out abnormal point feature extraction on each defect image in the textile product defect image set to obtain a defect feature extraction value; the region merging module is used for acquiring a defect extraction region set of the defect feature extraction value, and carrying out region merging on the defect extraction region set based on a textile product defect cluster analysis result to obtain a defect effective region set; the feature set determining module is used for determining a defect feature set of the textile product according to the defect feature extraction value and the defect effective area set; the model obtaining module is used for performing model training on the textile product defect feature set by using a deep learning network structure to obtain a textile product defect recognition model; and the detection and identification module is used for detecting and identifying the defect image of the textile product to be processed based on the textile product defect identification model.
One or more technical solutions provided in the present application have at least the following technical effects or advantages:
according to the method, the defect images of the textile product are acquired according to big data, a defect image set of the textile product is obtained, defect label classification is carried out on each defect image by constructing a defect label library of the textile product, defect label set information of the textile product is obtained according to classification results, image defects are subjected to cluster analysis to obtain a defect cluster analysis result of the textile product, further, abnormal feature extraction is carried out on each defect image, defect feature extraction values are obtained, after the defect extraction area sets of the defect feature extraction values are combined according to the defect cluster analysis result of the textile product, a defect effective area set is obtained, the defect feature set of the textile product is determined according to the defect feature extraction values, and after a defect recognition model of the textile product is obtained, the defect images of the textile product to be processed are detected and recognized. The method realizes the aim of efficiently identifying the defects of the textiles, achieves the technical effects of intelligently identifying the defects in the textiles and improving the identification efficiency and accuracy.
Drawings
For a clearer description of the technical solutions of the present application or of the prior art, the drawings used in the description of the embodiments or of the prior art will be briefly described below, it being obvious that the drawings in the description below are only exemplary and that other drawings can be obtained, without inventive effort, by a person skilled in the art from the drawings provided.
FIG. 1 is a schematic flow chart of a method for detecting defects in textile products according to an embodiment of the present application;
FIG. 2 is a schematic flow chart of obtaining extracted values of defect features in a method for detecting defects of textile products according to an embodiment of the present application;
FIG. 3 is a schematic flow chart of obtaining a standard set of images of defects of a textile product in a method for detecting defects of a textile product according to an embodiment of the present application;
FIG. 4 is a schematic diagram of a system for detecting defects in textile products according to the present application;
reference numerals illustrate: the device comprises an image acquisition module 11, a tag information acquisition module 12, a cluster analysis module 13, a feature extraction module 14, a region merging module 15, a feature set determination module 16, a model acquisition module 17 and a detection and identification module 18.
Description of the embodiments
The application solves the technical problems of lack of efficient detection of defects of textile products and low detection quality and efficiency in the prior art by providing the method and the system for detecting the defects of the textile products. The technical effects of intelligently identifying defects in textiles and improving identification efficiency and accuracy are achieved.
The data acquisition, storage, use, processing and the like in the technical scheme meet the relevant regulations of national laws and regulations.
In the following, the technical solutions in the present application will be clearly and completely described with reference to the accompanying drawings, and it should be understood that the described embodiments are only some embodiments of the present application, but not all embodiments of the present application, and the present application is not limited by the example embodiments described herein. All other embodiments, which can be made by one of ordinary skill in the art based on the embodiments of the present application without making any inventive effort, are intended to be within the scope of the present application. It should be further noted that, for convenience of description, only some, but not all of the drawings related to the present application are shown.
Examples
As shown in fig. 1, the present application provides a method for detecting defects in textile products, wherein the method comprises:
step S100: acquiring a textile product defect image set through big data acquisition;
further, after the collection of the defect images of the textile product is obtained through big data collection, step S100 in the embodiment of the present application further includes:
step S110: obtaining image transformation type information, wherein the image transformation type information comprises rotation, horizontal overturn, vertical overturn and scaling;
step S120: determining a defect image transformation coefficient according to the image transformation type information;
step S130: and carrying out data amplification on the textile product defect image set based on an image processing algorithm, and carrying out transformation output on the data amplification according to the defect image transformation coefficient to obtain an expanded textile product defect image set.
In particular, the image collection of defects of the textile product is that of collecting the image collection of defects in the textile product from Internet data and sensor data through big data. Alternatively, the collection method may be a web crawler or a website disclosure API, or the like. Because the devices for acquiring the images and the ports for uploading the images are different, the specifications and types of the defect images are various, and therefore, the images need to be subjected to image analysis. The image transformation type information is information of a specific operation type which needs to process and transform an image, and comprises the following steps: rotation, horizontal flip, vertical flip, zoom. The transformation coefficient of the defect image is determined according to the complexity of transforming the defect image determined by the image transformation type information. The image processing algorithm is an algorithm used for processing the image and comprises image stretching, shrinkage, rotation and the like. The step of performing data amplification on the textile product defect image set refers to performing transformation output on the textile product defect image set according to the defect image transformation coefficient, so that images in the image set are transformed to obtain the expanded textile product defect image set. Therefore, the aim of expanding an analysis database of the textile defect image is fulfilled, and the technical effects of reducing the non-uniformity of the image specification and improving the quality of the analysis image are achieved.
Step S200: constructing a textile product defect label library, and classifying the defect labels of the defect images based on the textile product defect label library to obtain textile product defect label set information;
specifically, the textile product defect label library is a label library obtained by marking different labels on different defect types and comprising defect types of textile products. And identifying and classifying the defects in the defect images through the defect tag library of the textile product, so as to classify the defect images in the image set and obtain the defect tag set information of the textile product. The textile product defect label set information is set information for characterizing all labels obtained after classifying defect marks in a textile product image. Therefore, the aim of identifying and classifying the defect types in the image set is fulfilled, and the technical effect of laying mats for the subsequent defect feature extraction is achieved.
Step S300: performing cluster analysis on the image defects according to the textile product defect label set information to obtain a textile product defect cluster analysis result;
specifically, the cluster analysis is to divide the image defects of the same label into the same group according to different labels to which the image defects belong, and the image defects of different labels are divided into different groups. And the textile product defect cluster analysis result is that the defects in the textile product defect image set are classified according to labels to obtain classified image combination. Therefore, the aim of improving analysis efficiency is fulfilled, and the technical effect of improving the intuitiveness of image classification is achieved.
Step S400: performing abnormal point feature extraction on each defect image in the defect image set of the textile product to obtain a defect feature extraction value;
further, as shown in fig. 2, the obtaining the extracted value of the defect feature, step S400 in the embodiment of the present application further includes:
step S410: performing image preprocessing on the textile product defect image set to obtain a standard textile product defect image set;
step S420: gray value display is carried out on the basis of each defect image in the standard textile product defect image set, and textile product defect image gray value information is obtained;
step S430: constructing an image gray level distribution coordinate system, and mapping each defect image into the image gray level distribution coordinate system to obtain a defect image gray level histogram;
step S440: calculating a gray average value based on the gray histogram of the defect image to obtain gray average value information of the defect image;
step S450: and calculating gray standard deviation of the gray value information of the defect image of the textile product and the gray mean value information of the defect image, and taking the gray standard deviation as a defect characteristic extraction value.
Further, as shown in fig. 3, the performing image preprocessing on the set of defect images of the textile product to obtain a set of defect images of the standard textile product, step S410 of the embodiment of the present application includes:
step S411: denoising the textile product defect image set based on a filtering algorithm to obtain a denoised textile product defect image set;
step S412: carrying out grey conversion on the denoised textile product defect image set to obtain a textile product defect grey image set;
step S413: and carrying out image enhancement on the textile product defect gray image set based on the generation of an antagonism network to obtain the standard textile product defect image set.
Specifically, the image preprocessing is used for eliminating irrelevant information in the image and enhancing the detectability of the information. The images in the collection of defective images of the textile product can generate various noises during the digitizing and transmitting processes, which are represented as isolated pixels or pixel blocks causing a strong visual effect, disturbing the observable information of the images. The filtering algorithm is to perform weighted average on a plurality of pixel point values in a noise point neighborhood in an image in the textile product defect image set, so that the purpose of removing noise from the image is achieved. The image collection of the defects of the denoised textile product is an image collection of the clearly detectable image information subjected to noise removal.
Wherein the graying conversion is by converting the images in the image set into a new gray value according to a mathematical formula, optionally the graying conversion method comprises: logarithmic transformation, exponential transformation, gamma transformation, histogram equalization, etc. And forming an image with only one sampling color per pixel by adjusting the gray value in the image, so as to form the textile product defect gray image set. The generating countermeasure network is a deep learning model for performing unsupervised learning, and high-quality output is generated through mutual game learning of the generating model and the judging model. Image enhancement is carried out on the image set through the generation countermeasure network, the low-resolution image is generated in the first stage, defects in the low-resolution image in the first stage are corrected in the second stage, image details are perfected, and therefore a high-resolution image is generated. Therefore, the aim of preprocessing the image to obtain a standard textile product defect image set is fulfilled, and the technical effects of improving the image quality and enhancing the identifiable information of the image are achieved.
Specifically, each defect image includes not only the image of the defect but also the image of the normal point of the textile product, so that the feature extraction is performed on the abnormal point, namely the defect, in the defect image, so as to prepare for the subsequent determination of the defect range. The gray value information of the defect image of the textile product is gray value information of each pixel point of each defect image obtained by displaying the gray value of each defect image. The image gray level distribution coordinate system is used for reflecting the occurrence condition of different gray level values of the image, the abscissa represents each gray level value of the image, and the ordinate represents the number of each gray level value of the image. The gray level histogram of each defect image can be obtained by mapping each defect image into the gray level distribution coordinate system of the image, wherein the gray level histogram of each defect image reflects the distribution condition of each gray level value in the image. Further, the average gray level calculation is performed to obtain the arc average value information of the defect image reflecting the average gray level of the defect image. And calculating the gray value information and the gray average value information of each image to obtain the gray standard deviation of the dispersion degree of the gray value and the average value of the reflected image. The gray standard deviation can reflect the deviation of gray values at the defects and the gray values of normal spinning points, and can be used as the defect characteristic extraction value. Therefore, the determination of the feature extraction target in the process of identifying the defects is realized, and the technical effects of improving the extraction efficiency and the extraction accuracy are achieved.
Step S500: acquiring a defect extraction area set of the defect feature extraction value, and carrying out area combination on the defect extraction area set based on a textile product defect cluster analysis result to obtain a defect effective area set;
step S600: determining a defect feature set of the textile product according to the defect feature extraction value and the defect effective area set;
specifically, the defect extraction area set is a defect area set determined by the defect feature extraction value. And carrying out region merging on the regions where the defects belonging to the same class are located based on the textile product defect cluster analysis result, thereby obtaining the defect effective region set. Wherein the defect effective area set is a set integrating two aspects of defect types and defect areas, and the defect feature set of the textile product is determined by combining the defect feature extraction value. Wherein the set of textile product defect features is a set of features that reflects the textile product defect features and the defective area in which it is located. Therefore, the technical effect of providing the characteristics for the defect characteristic recognition of the follow-up textile products is achieved.
Step S700: performing model training on the textile product defect feature set by using a deep learning network structure to obtain a textile product defect recognition model;
further, the performing model training on the textile product defect feature set by using the deep learning network structure to obtain a textile product defect recognition model, and step S700 of the embodiment of the present application further includes:
step S710: identifying the defect feature set of the textile product, and dividing the identified feature data set into a training set, a verification set and a test set;
step S720: performing supervision training on the deep learning network structure based on the training set to obtain an initial product defect recognition model;
step S730: and verifying and testing the initial product defect recognition model based on the verification set and the test set until the model recognition accuracy reaches a preset accuracy, and obtaining the textile product defect recognition model.
Further, step S700 in the embodiment of the present application further includes:
step S740: performing environmental interference analysis on the defect image of the textile product to be processed to obtain a detection environmental interference factor;
step S750: determining a loss weight impact factor based on the detected environmental interference factor;
step S760: and iteratively updating the textile product defect recognition model according to the loss weight influence factor.
Specifically, the progress of training during model training can be achieved by identifying the textile product defect feature set, and the identified feature data set is further divided into a training set, a verification set and a test set, and optionally, the division method can be used for dividing according to the ratio of 4:3:3. The training set is used for performing supervision training on the deep learning network, so that the network layer can identify the learned defect characteristics, when the defect characteristics are encountered again, the types and the areas of the defects can be identified, the model obtained after learning is the initial product defect identification model, and the defect characteristics can be identified by the model at the moment. And further, verifying the accuracy of the initial product defect recognition model through the verification set, and testing the recognition speed of the initial product defect recognition model through the test set, wherein the recognition accuracy of the model can reach the preset accuracy. The preset accuracy is the accuracy of the model which is preset and can reach the accuracy of production and use.
Specifically, the environmental interference is analyzed, the display condition of the defect image of the textile product to be processed is analyzed, and the detected environmental interference factor is a factor affecting the display quality of the image and can be: light, sampling rate, resolution, etc. The loss weight impact factor is obtained by determining the impact degree of the interference factor. The loss weight influence factor is quantized data used for representing the influence degree of the interference factor on image recognition. The loss weight influence factors are used for carrying out iterative updating on the textile product defect identification model, so that the loss caused by the influence factors is compensated, the aim of optimizing and updating the model is fulfilled, and the technical effect of improving the accuracy of model identification is achieved.
Step S800: and detecting and identifying the defect image of the textile product to be treated based on the defect identification model of the textile product.
Further, step S800 in the embodiment of the present application further includes:
step S810: the quality of the textile product is graded according to the customer demand standard, and the quality grade of the textile product is obtained;
step S820: if the quality grade of the textile product does not reach the preset quality grade, counting the defect detection result of the textile product to obtain a defect detection type counting result and a defect detection frequency counting result;
step S830: obtaining a defect generation influence parameter according to the defect detection type statistical result and the defect detection frequency statistical result;
step S840: and carrying out parameter optimization adjustment on the textile product production process based on the defect generation influence parameters.
Specifically, the defect type can be identified by inputting the defect image of the textile product to be treated into the defect identification model of the textile product to carry out defect detection and identification, and identifying the defect condition of the textile product to be treated.
Specifically, the customer requirement standard refers to a standard established for meeting the requirements of customers, and the requirements of products must be met. And obtaining the quality grade of the textile product by evaluating the degree of satisfaction of the standard according to the customer demand standard. Wherein the textile product quality grade may be a premium, a first grade, a pass, and a fail. The preset quality grade is a preset grade meeting the quality requirement of the customer. The defect detection type statistics are specific types of defects detected from the textile. The defect detection frequency is the frequency with which different defect types occur. The defect generation influence parameter is a parameter which reflects the degree of influence of the type and frequency of defect generation on the quality condition of the textile product. The production process of the textile product is adjusted according to the defect generation influence parameters, and the quality of the textile product is improved by adjusting the process and reducing the formation of defects. Therefore, the aim of adjusting the process based on the detection result is fulfilled, and the technical effect of improving the quality of textile products is achieved.
In summary, the method for detecting defects of textile products provided by the application has the following technical effects:
1. the method comprises the steps of collecting defect images of textile products based on big data, obtaining a defect image set of the textile products, classifying defect labels of the defect images, obtaining defect label set information of the textile products according to classification results, obtaining a defect cluster analysis result of the textile products through cluster analysis, extracting abnormal point features in the defect images, carrying out region merging on defect extraction region sets of defect feature extraction values according to the cluster analysis result to obtain a defect effective region set, determining a defect feature set of the textile products according to the defect feature extraction values, and detecting and identifying the defect images of the textile products to be processed after a defect identification model of the textile products is obtained. The method realizes the aim of efficiently identifying the defects of the textiles, achieves the technical effects of intelligently identifying the defects in the textiles and improving the identification efficiency and accuracy.
2. According to the method, the image preprocessing is carried out on the textile product defect image set to obtain the standard textile product defect image set, gray value display is carried out on each defect image in the set to obtain the textile product defect image gray value information, the image gray distribution coordinate system is constructed, each defect image is mapped into the image gray distribution coordinate system to obtain the defect image gray histogram, gray average value calculation is carried out to obtain the defect image gray average value information, gray standard deviation of the textile product defect image gray value information and the defect image gray average value information is calculated, and the gray standard deviation is used as a defect feature extraction value. Therefore, the aim of preprocessing the image to obtain a standard textile product defect image set is fulfilled, and the technical effects of improving the image quality and enhancing the identifiable information of the image are achieved.
Examples
Based on the same inventive concept as in the previous embodiment for a method for detecting defects of textile products, as shown in fig. 4, the present application also provides a system for detecting defects of textile products, wherein the system comprises:
the image acquisition module 11 is used for acquiring a textile product defect image set through big data acquisition;
the label information obtaining module 12 is used for constructing a textile product defect label library, classifying the defect labels of the defect images based on the textile product defect label library, and obtaining textile product defect label set information;
the cluster analysis module 13 is used for carrying out cluster analysis on the image defects according to the textile product defect label set information to obtain a textile product defect cluster analysis result;
the feature extraction module 14 is used for extracting abnormal point features of each defect image in the defect image set of the textile product to obtain a defect feature extraction value;
the region merging module 15 is used for obtaining a defect extraction region set of the defect feature extraction value, and carrying out region merging on the defect extraction region set based on a textile product defect cluster analysis result to obtain a defect effective region set;
a feature set determining module 16, wherein the feature set determining module 16 is configured to determine a defect feature set of the textile product according to the defect feature extraction value and the defect effective area set;
the model obtaining module 17 is used for performing model training on the textile product defect feature set by using a deep learning network structure to obtain a textile product defect recognition model;
the detection and identification module 18 is used for detecting and identifying the defect image of the textile product to be processed based on the defect identification model of the textile product, wherein the detection and identification module 18 is used for detecting and identifying the defect image of the textile product to be processed.
Further, the system further comprises:
the standard image processing unit is used for carrying out image preprocessing on the textile product defect image set to obtain a standard textile product defect image set;
the gray value information acquisition unit is used for displaying gray values based on each defect image in the standard textile product defect image set to obtain textile product defect image gray value information;
the histogram obtaining unit is used for constructing an image gray level distribution coordinate system, mapping each defect image into the image gray level distribution coordinate system and obtaining a defect image gray level histogram;
the gray average value obtaining unit is used for carrying out gray average value calculation based on the gray histogram of the defect image to obtain gray average value information of the defect image;
the standard deviation obtaining unit is used for calculating gray standard deviation of the gray value information of the defect image of the textile product and the gray average value information of the defect image, and taking the gray standard deviation as a defect characteristic extraction value.
Further, the system further comprises:
the denoising processing unit is used for denoising the textile product defect image set based on a filtering algorithm to obtain a denoised textile product defect image set;
the grey level conversion unit is used for carrying out grey level conversion on the denoised textile product defect image set to obtain a textile product defect grey level image set;
and the image enhancement unit is used for carrying out image enhancement on the textile product defect gray image set based on the generation of an antagonism network to obtain the standard textile product defect image set.
Further, the system further comprises:
a transformation type obtaining unit for obtaining image transformation type information including rotation, horizontal flip, vertical flip, and scaling;
a transformation coefficient obtaining unit for determining a transformation coefficient of the defective image according to the image transformation type information;
the data amplification unit is used for carrying out data amplification on the textile product defect image set based on an image processing algorithm, and the data amplification is converted and output according to the defect image conversion coefficient to obtain an expanded textile product defect image set.
Further, the system further comprises:
the identification unit is used for identifying the defect feature set of the textile product and dividing the identified feature data set into a training set, a verification set and a test set;
the training unit is used for performing supervision training on the deep learning network structure based on the training set to obtain an initial product defect recognition model;
and the verification test unit is used for verifying and testing the initial product defect recognition model based on the verification set and the test set until the model recognition accuracy reaches a preset accuracy to obtain the textile product defect recognition model.
Further, the system further comprises:
the interference analysis unit is used for carrying out environmental interference analysis on the to-be-processed textile product defect image to obtain a detection environmental interference factor;
an influence factor determining unit configured to determine a loss weight influence factor based on the detection environment interference factor;
and the iteration updating unit is used for carrying out iteration updating on the textile product defect identification model according to the loss weight influence factor.
Further, the system further comprises:
the quality rating unit is used for carrying out quality rating on the textile product according to the customer demand standard to obtain the quality grade of the textile product;
the statistics unit is used for counting the defect detection results of the textile products if the quality grade of the textile products does not reach the preset quality grade, and obtaining the defect detection type statistics results and the defect detection frequency statistics results;
the influence parameter obtaining unit is used for obtaining a defect generation influence parameter according to the defect detection type statistical result and the defect detection frequency statistical result;
and the optimizing and adjusting unit is used for optimizing and adjusting parameters of the textile product production process based on the defect generation influence parameters.
The embodiments described in the present specification are described in a progressive manner, and each embodiment focuses on the difference from the other embodiments, and a method and a specific example for detecting defects of a textile product in the first embodiment of fig. 1 are equally applicable to a system for detecting defects of a textile product in the first embodiment, and a system for detecting defects of a textile product in the first embodiment will be apparent to those skilled in the art from the foregoing detailed description of a method for detecting defects of a textile product, so that the detailed description of the system for detecting defects of a textile product in the first embodiment is omitted herein for brevity of the present specification. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant points refer to the description of the method section.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the application. Thus, the present application is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (8)
1. A method for defect detection in textile products, the method comprising:
acquiring a textile product defect image set through big data acquisition;
constructing a textile product defect label library, and classifying the defect labels of the defect images based on the textile product defect label library to obtain textile product defect label set information;
performing cluster analysis on the image defects according to the textile product defect label set information to obtain a textile product defect cluster analysis result;
performing abnormal point feature extraction on each defect image in the defect image set of the textile product to obtain a defect feature extraction value;
acquiring a defect extraction area set of the defect feature extraction value, and carrying out area combination on the defect extraction area set based on a textile product defect cluster analysis result to obtain a defect effective area set;
determining a defect feature set of the textile product according to the defect feature extraction value and the defect effective area set;
performing model training on the textile product defect feature set by using a deep learning network structure to obtain a textile product defect recognition model;
and detecting and identifying the defect image of the textile product to be treated based on the defect identification model of the textile product.
2. A method as in claim 1 wherein said obtaining a defect signature extraction value comprises:
performing image preprocessing on the textile product defect image set to obtain a standard textile product defect image set;
gray value display is carried out on the basis of each defect image in the standard textile product defect image set, and textile product defect image gray value information is obtained;
constructing an image gray level distribution coordinate system, and mapping each defect image into the image gray level distribution coordinate system to obtain a defect image gray level histogram;
calculating a gray average value based on the gray histogram of the defect image to obtain gray average value information of the defect image;
and calculating gray standard deviation of the gray value information of the defect image of the textile product and the gray mean value information of the defect image, and taking the gray standard deviation as a defect characteristic extraction value.
3. A method according to claim 2, wherein said image preprocessing said set of textile product defect images to obtain a standard set of textile product defect images comprises:
denoising the textile product defect image set based on a filtering algorithm to obtain a denoised textile product defect image set;
carrying out grey conversion on the denoised textile product defect image set to obtain a textile product defect grey image set;
and carrying out image enhancement on the textile product defect gray image set based on the generation of an antagonism network to obtain the standard textile product defect image set.
4. A method according to claim 1, wherein said acquiring a collection of images of defects of textile products by big data acquisition comprises:
obtaining image transformation type information, wherein the image transformation type information comprises rotation, horizontal overturn, vertical overturn and scaling;
determining a defect image transformation coefficient according to the image transformation type information;
and carrying out data amplification on the textile product defect image set based on an image processing algorithm, and carrying out transformation output on the data amplification according to the defect image transformation coefficient to obtain an expanded textile product defect image set.
5. The method of claim 1, wherein said model training said set of textile product defect features using a deep learning network architecture to obtain a textile product defect identification model, comprising:
identifying the defect feature set of the textile product, and dividing the identified feature data set into a training set, a verification set and a test set;
performing supervision training on the deep learning network structure based on the training set to obtain an initial product defect recognition model;
and verifying and testing the initial product defect recognition model based on the verification set and the test set until the model recognition accuracy reaches a preset accuracy, and obtaining the textile product defect recognition model.
6. The method of claim 1, wherein the method comprises:
performing environmental interference analysis on the defect image of the textile product to be processed to obtain a detection environmental interference factor;
determining a loss weight impact factor based on the detected environmental interference factor;
and iteratively updating the textile product defect recognition model according to the loss weight influence factor.
7. The method of claim 1, wherein the method comprises:
the quality of the textile product is graded according to the customer demand standard, and the quality grade of the textile product is obtained;
if the quality grade of the textile product does not reach the preset quality grade, counting the defect detection result of the textile product to obtain a defect detection type counting result and a defect detection frequency counting result;
obtaining a defect generation influence parameter according to the defect detection type statistical result and the defect detection frequency statistical result;
and carrying out parameter optimization adjustment on the textile product production process based on the defect generation influence parameters.
8. A system for defect detection of textile products, the system comprising:
the image acquisition module is used for acquiring a textile product defect image set through big data acquisition;
the label information obtaining module is used for constructing a textile product defect label library, classifying the defect labels of the defect images based on the textile product defect label library, and obtaining textile product defect label set information;
the cluster analysis module is used for carrying out cluster analysis on the image defects according to the textile product defect label set information to obtain a textile product defect cluster analysis result;
the feature extraction module is used for carrying out abnormal point feature extraction on each defect image in the textile product defect image set to obtain a defect feature extraction value;
the region merging module is used for acquiring a defect extraction region set of the defect feature extraction value, and carrying out region merging on the defect extraction region set based on a textile product defect cluster analysis result to obtain a defect effective region set;
the feature set determining module is used for determining a defect feature set of the textile product according to the defect feature extraction value and the defect effective area set;
the model obtaining module is used for performing model training on the textile product defect feature set by using a deep learning network structure to obtain a textile product defect recognition model;
and the detection and identification module is used for detecting and identifying the defect image of the textile product to be processed based on the textile product defect identification model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310298187.7A CN116309493A (en) | 2023-03-24 | 2023-03-24 | Method and system for detecting defects of textile products |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310298187.7A CN116309493A (en) | 2023-03-24 | 2023-03-24 | Method and system for detecting defects of textile products |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116309493A true CN116309493A (en) | 2023-06-23 |
Family
ID=86828500
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310298187.7A Pending CN116309493A (en) | 2023-03-24 | 2023-03-24 | Method and system for detecting defects of textile products |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116309493A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117115147A (en) * | 2023-10-19 | 2023-11-24 | 山东华盛创新纺织科技有限公司 | Textile detection method and system based on machine vision |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180322623A1 (en) * | 2017-05-08 | 2018-11-08 | Aquifi, Inc. | Systems and methods for inspection and defect detection using 3-d scanning |
| US20190333208A1 (en) * | 2018-04-25 | 2019-10-31 | Applied Materials Israel, Ltd. | Method of defect classification and system thereof |
| CN110490874A (en) * | 2019-09-04 | 2019-11-22 | 河海大学常州校区 | Weaving cloth surface flaw detecting method based on YOLO neural network |
| CN114757900A (en) * | 2022-03-31 | 2022-07-15 | 启东新朋莱纺织科技有限公司 | Artificial intelligence-based textile defect type identification method |
| CN114998695A (en) * | 2022-07-18 | 2022-09-02 | 深圳市前海泽金产融科技有限公司 | Method and system for improving image recognition speed |
| CN115294136A (en) * | 2022-10-09 | 2022-11-04 | 播种者工业智能科技(苏州)有限公司 | Artificial intelligence-based construction and detection method for textile fabric flaw detection model |
-
2023
- 2023-03-24 CN CN202310298187.7A patent/CN116309493A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180322623A1 (en) * | 2017-05-08 | 2018-11-08 | Aquifi, Inc. | Systems and methods for inspection and defect detection using 3-d scanning |
| US20190333208A1 (en) * | 2018-04-25 | 2019-10-31 | Applied Materials Israel, Ltd. | Method of defect classification and system thereof |
| CN110490874A (en) * | 2019-09-04 | 2019-11-22 | 河海大学常州校区 | Weaving cloth surface flaw detecting method based on YOLO neural network |
| CN114757900A (en) * | 2022-03-31 | 2022-07-15 | 启东新朋莱纺织科技有限公司 | Artificial intelligence-based textile defect type identification method |
| CN114998695A (en) * | 2022-07-18 | 2022-09-02 | 深圳市前海泽金产融科技有限公司 | Method and system for improving image recognition speed |
| CN115294136A (en) * | 2022-10-09 | 2022-11-04 | 播种者工业智能科技(苏州)有限公司 | Artificial intelligence-based construction and detection method for textile fabric flaw detection model |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117115147A (en) * | 2023-10-19 | 2023-11-24 | 山东华盛创新纺织科技有限公司 | Textile detection method and system based on machine vision |
| CN117115147B (en) * | 2023-10-19 | 2024-01-26 | 山东华盛创新纺织科技有限公司 | Textile detection method and system based on machine vision |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114549522B (en) | Textile quality detection method based on target detection | |
| CN110148130B (en) | Method and device for detecting part defects | |
| CN111383209B (en) | Unsupervised flaw detection method based on full convolution self-encoder network | |
| CN116758491B (en) | Printing monitoring image analysis method and system applied to 3D printing | |
| CN107369155A (en) | A kind of cloth surface defect detection method and its system based on machine vision | |
| CN110188806A (en) | A kind of large circle machine fabric defects detection and classification method based on machine vision | |
| Fekri-Ershad et al. | A robust approach for surface defect detection based on one dimensional local binary patterns | |
| CN114782329A (en) | Bearing defect damage degree evaluation method and system based on image processing | |
| CN111724376B (en) | Paper disease detection method based on texture feature analysis | |
| CN111402236A (en) | Hot-rolled strip steel surface defect grading method based on image gray value | |
| CN116309493A (en) | Method and system for detecting defects of textile products | |
| CN115018790A (en) | Workpiece surface defect detection method based on anomaly detection | |
| CN110458809B (en) | Yarn evenness detection method based on sub-pixel edge detection | |
| CN115018785A (en) | Hoisting steel wire rope tension detection method based on visual vibration frequency identification | |
| CN117237747B (en) | Hardware defect classification and identification method based on artificial intelligence | |
| Zhang et al. | Fabric defect detection based on visual saliency map and SVM | |
| CN117315365A (en) | Camshaft surface damage detecting system based on visual analysis | |
| CN116735610A (en) | Steel pipe surface defect detection method based on machine vision | |
| Jeffrey Kuo et al. | Self-organizing map network for automatically recognizing color texture fabric nature | |
| Guo et al. | Fault diagnosis of power equipment based on infrared image analysis | |
| CN113077002A (en) | Machine olfaction visualization sensing data analysis method based on spatial heterodyne Raman spectrum | |
| Hu et al. | A lightweight reconstruction network for surface defect inspection | |
| CN112508946A (en) | Cable tunnel abnormity detection method based on antagonistic neural network | |
| CN111696070A (en) | Multispectral image fusion power internet of things fault point detection method based on deep learning | |
| CN118379283B (en) | Flat wire motor stator surface defect detection method, device, equipment and 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 |