CN113920112A - Fabric flaw detection method based on independent classification type feature extraction - Google Patents
Fabric flaw detection method based on independent classification type feature extraction Download PDFInfo
- Publication number
- CN113920112A CN113920112A CN202111353364.4A CN202111353364A CN113920112A CN 113920112 A CN113920112 A CN 113920112A CN 202111353364 A CN202111353364 A CN 202111353364A CN 113920112 A CN113920112 A CN 113920112A
- Authority
- CN
- China
- Prior art keywords
- fabric
- color
- image
- feature extraction
- color block
- 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
- 238000001514 detection method Methods 0.000 title claims abstract description 73
- 239000004744 fabric Substances 0.000 title claims abstract description 61
- 238000000605 extraction Methods 0.000 title claims abstract description 35
- 230000007547 defect Effects 0.000 claims abstract description 59
- 238000000034 method Methods 0.000 claims abstract description 57
- 230000002159 abnormal effect Effects 0.000 claims abstract description 12
- 239000003086 colorant Substances 0.000 claims abstract description 4
- 238000007781 pre-processing Methods 0.000 claims description 14
- 238000012545 processing Methods 0.000 claims description 7
- 238000001914 filtration Methods 0.000 claims description 6
- 238000005070 sampling Methods 0.000 claims description 4
- 230000003628 erosive effect Effects 0.000 claims description 3
- 238000006243 chemical reaction Methods 0.000 claims description 2
- 230000002950 deficient Effects 0.000 claims 2
- 230000007797 corrosion Effects 0.000 claims 1
- 238000005260 corrosion Methods 0.000 claims 1
- 238000005516 engineering process Methods 0.000 abstract description 3
- 239000000284 extract Substances 0.000 abstract description 3
- 238000004458 analytical method Methods 0.000 description 11
- 238000010586 diagram Methods 0.000 description 9
- 239000004753 textile Substances 0.000 description 9
- 238000011160 research Methods 0.000 description 8
- 230000005856 abnormality Effects 0.000 description 3
- 238000013461 design Methods 0.000 description 3
- 238000011161 development Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000002372 labelling Methods 0.000 description 3
- 238000004519 manufacturing process Methods 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000004927 fusion Effects 0.000 description 2
- 238000005286 illumination Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 230000004308 accommodation Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000004438 eyesight Effects 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000012805 post-processing Methods 0.000 description 1
- 238000011268 retreatment Methods 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Quality & Reliability (AREA)
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a fabric flaw detection method based on independent classification type feature extraction, which mainly aims at flaw detection of fabrics containing discrete regular patterns or plain colors, obtains a model containing color block type information by manually modeling normal fabrics, then divides color blocks and marks abnormal color blocks according to the matching degree of the fabric color blocks and the model under a certain threshold value, then processes corresponding images on the divided color blocks, and respectively extracts and marks corresponding flaws by using convex hull defects and embedded defect detection algorithms so as to realize independent classification type extraction of the fabric flaws on the characteristics of chroma, convex hull, embedded color and the like. The fabric detection method provided by the invention comprises the steps of building a fabric model, designing a detection principle and the like, and combines a multithreading technology to realize the rapid extraction of independent classification of fabric flaws, so that the fabric flaw detection method can be more conveniently and efficiently applied to the fabric flaw detection process.
Description
Technical Field
The invention belongs to the field of computer vision image processing, relates to a method for quickly extracting fabric surface flaws, and particularly relates to a fabric flaw detection method based on independent classification type feature extraction.
Background
The textile industry is the backbone of our economy, and its development is closely related to our daily lives. With the great development of internationalization, the export quantity of textiles in China is increased greatly, but the problems brought by the export quantity of the textiles are gradually brought to the front of people. In the large environment of textile industry, machines with textile functions have appeared for a long time, so that both hands of people are greatly liberated, but at present, a great deal of manual investment still exists in the field of textile quality detection, and therefore, how to improve the quality of textile products and the production efficiency is still one of very important problems.
With the continuous development of image processing technology, some detection means correspondingly appear for detecting textile fabric flaws, so that certain possibility exists for the automation of the detection of the textile fabric flaws, especially advanced algorithms are proposed in recent years, and the precision and the breadth of flaw detection are always important for the research in the field, regardless of principle or performance, compared with the initial stage. In spite of the current research situation at home and abroad, a statistical-based method, a frequency-domain-based method and a model-based method are available in texture feature extraction, and the methods have a certain effect on detecting fabric flaws to a certain extent, but most of the methods are only limited to the test stage, and only aim at plain fabrics, so that the methods can be really applied to the actual working environment and are few, and the field has a large research gap.
Compared with the defect detection of printed fabrics, the defect detection of the plain fabrics is relatively easy in algorithm implementation due to single color, and is also the key point of algorithm research of broad scholars. For example, Conci et al use the parting idea for fabric flaw detection and adopt a differential box counting method to obtain faster detection efficiency; henbury et al use morphology to obtain image direction features for fabric flaw detection with good results but with manual intervention. Therefore, at present, a great deal of improvement space still exists in fabric flaw detection, particularly in flaw detection research of printed fabrics. In view of the above, the present invention provides a fabric defect detection method based on independent classification type feature extraction, which extracts defects in an independent classification type by using a multithreading technology through different image preprocessing modes to realize defect detection on fabrics containing regular patterns or plain colors.
Disclosure of Invention
Aiming at the defects or shortcomings existing in the fabric flaw detection process, the invention aims to provide a fabric flaw detection method based on independent classification type feature extraction, and solves the problems that the flaw detection object is single, the detection efficiency is low, and the actual production is difficult to apply.
In order to achieve the above purpose, the technical scheme of the invention is as follows: the method comprises a modeling process of a sample image, a preprocessing process of a detection image, a color block matching and extracting process, and a process of performing independent classification type feature extraction by using algorithms such as embedded detection and convex hull detection.
The sample image modeling process comprises seed point selection, color block clustering and color block information acquisition, wherein the seed point selection is to randomly select a pixel point on a certain color block of the sample image, and the chroma distance colorDis value is calculated by
To calculate in the formula
The respective accumulated mean values of the color blocks are calculated by
Calculated, we can get the following conditions according to the following: v is less than or equal to color DisthresholdTo determine whether two adjacent elements belong to the color block class, wherein vthresholdIs a distance threshold. By analogy, each type of color block on the sample image is clustered correspondingly, and corresponding color block clustering information can be obtained, wherein the color block clustering information comprises an average LAB value of the color block and a corresponding LAB valueThe number of color block types and the like;
the detection image preprocessing mainly relates to the links of down sampling, chrominance space conversion, image Gaussian filtering and the like. In the downsampling, in order to reduce the complexity of image processing, a scale factor is set to be 0.5, and the size of an image is only half of that of an original image at the moment; in addition, the image channel is converted from an RGB space to an LAB space, so that the extraction of a color block at the back is easier; and then, Gaussian filtering is carried out, and the size of a convolution kernel of the Gaussian filtering selected here is 7 x 7, so that the influence of noise on the extraction of the subsequent color blocks is reduced;
the sample image is also required to be subjected to the same preprocessing process as the detection image before the model is established, but the processing of down sampling is lacked, so that the purpose of ensuring that the model is more appropriate to the actual detection process is achieved, and the detection performance is prevented from being reduced;
the distance threshold depends on the illumination intensity of image acquisition, and the parameter needs to be calibrated under experimental conditions so as to ensure the accuracy and the effectiveness of the model; the threshold value determines the effect of extracting the subsequent color blocks, if the value is set to be too small, the color blocks are separated too discretely, and in severe cases, the contour of the regular color blocks is incomplete, so that the contour analysis of the color blocks cannot be carried out, and if the value is too large, some color blocks are fused and are not beneficial to carrying out the contour analysis; therefore, the selection of the distance threshold is very important, and the field calibration is necessary;
the color block contour defect is caused by incomplete extraction of regular color blocks due to illumination of a detection image or a camera when the regular color blocks on the surface of the fabric are extracted;
the color lump fusion is that the physical positions of the scattered regular color lumps on the image are close, and the fusion phenomenon of the color lumps is extracted under the condition of excessively large distance from the threshold value, so that the color lump outline is deformed, and the outline analysis process is influenced;
color block matching and extracting, namely comparing the information of each pixel point of a detected image with various color blocks of a model, thereby performing information matching, marking a region which does not belong to the color block of the model on a defect marking map, independently extracting and recording each color block on a plurality of new maps, then performing corresponding retreatment on the new maps, respectively and simultaneously sending the new maps into an embedded and convex hull detection algorithm for fabric defect extraction in a multi-thread program design mode, and finally respectively marking two types of defects of the embedded and convex hulls on the defect marking map;
the convex hull flaws refer to flaws which change the edge shape of the color block of the original model;
the embedded flaws refer to flaws existing in the color blocks in the model under the condition of normal chromaticity, which do not affect the outline shape of the color blocks but affect the position relation of the outlines of the color blocks;
the embedded and convex hull defect detection process comprises the analysis of the outline shape and the position relation, mainly aims to solve the problems of outline shape deformation and outline position relation abnormity caused by the defects of color blocks, and aims to extract the fabric defects caused by the two conditions;
the contour shape abnormity is to extract the contour of each extracted color block, and to use convex hull analysis to find the area of the regular contour edge where the convex hull is generated due to the flaw. Because the contour of the extracted color block is not completely smooth, a corresponding threshold value is set to filter some false-detected convex hull areas;
furthermore, the abnormal profile relation is to judge whether a profile has an outer profile to determine whether the profile belongs to an embedded flaw, and a certain threshold value is also required to be set to filter out an excessively small embedded profile to avoid false detection.
Compared with the prior art, the invention has the beneficial technical effects that:
1. the invention provides a fabric flaw detection method based on independent classification type feature extraction, which can solve the problem of detecting basic printed fabric flaws to a certain extent, and can greatly improve the efficiency of fabric flaw detection when being applied to the actual work of a cloth inspection machine, particularly on some printed fabric detections. The method comprises the steps of modeling a sample image, matching the sample image with a detection image, marking an area with abnormal color intensity, extracting corresponding color blocks, respectively sending the color blocks into corresponding detection algorithms under the condition of multithread design, and independently extracting fabric flaws in types. The complexity of fabric flaw detection is optimized to a great extent in the process, particularly in algorithm arrangement, modular program design can be achieved, independent extraction can be performed on the fabric flaw detection according to flaw features, and the phenomenon of false detection and missed detection caused by aliasing of features among various types of flaws is avoided;
2. although a great deal of research results exist in the research on the fabric flaw detection algorithm at present, most of the research results only can stay in a test stage, are unstable or single in detection object, or have high time or space complexity, and have some limitations in practical production. The method analyzes the reasons for the generation of the flaws and extracts the flaws in an independent type-classifying mode, so that on one hand, the complexity of an algorithm is reduced, and on the other hand, the fabric flaw detection process is more orderly and complete, and the precision and the breadth of the fabric flaw detection are realized;
drawings
FIG. 1 is a flow chart of a method of flaw feature extraction;
FIG. 2 is a sample graph of flaw types;
FIG. 3 is a diagram of an in-line defect detection process;
FIG. 4 is a diagram of a convex hull defect detection process;
FIG. 5 is a schematic diagram of a flaw independent classification extraction process;
FIG. 6 is a flowchart illustrating the embedded, convex hull defect detection of the present invention;
wherein in FIG. 2, 1-embedded defect; 2-colorimetric abnormalities; 3-convex hull defect;
in FIG. 5, 4-Defect extraction image; 5-flaw labeling image.
Detailed Description
The present invention is described in further detail below with reference to the attached drawings.
As shown in fig. 1, the invention provides a fabric defect detection method based on independent classification type feature extraction, which comprises a preprocessing method and a modeling principle for establishing a model process, a principle for a chromaticity anomaly detection process, a method for color block extraction, design of convex hull and embedded algorithm, and a method for summarizing and marking fabric defects;
as shown in fig. 2, a sample diagram of the types of defects is shown, in which an in-line type defect 1, a chroma abnormal defect 2 and a convex hull type defect 3 are included, and these types of defects are three types that are often found in a fabric of a discrete regular pattern type; the chroma abnormal defect 2 is a type of defect which is judged as a defect because the color of one area is not in the model, the embedded type defect 1 is a type of defect which is marked as a defect because the position inclusion relation is wrong although the color of one area is included in the model, and the convex hull type defect 3 is a type of defect which is marked as a defect because the color of one area is also included in the model and the shape of a color block edge is influenced;
the color lump extraction process mainly analyzes each pixel point of the detected image according to the information in the model, thereby marking the color lump category;
as shown in fig. 3 and 4, which are respectively an embedded defect detection process diagram and a convex hull defect detection process diagram, wherein the embedded defect detection process diagram shows how an embedded detection algorithm analyzes the contour relationship of a color block, and the convex hull detection process diagram shows how a convex hull detection algorithm analyzes the contour shape of the color block, and the specific algorithm implementation principle is shown in fig. 6;
the embedded detection algorithm and the convex hull detection algorithm are arranged on two threads in parallel to ensure the efficiency of detecting the fabric flaws;
as shown in fig. 6, the flowchart of the defect detection for embedded and convex hulls of the present invention mainly includes the processes of image erosion, contour extraction, contour analysis, marking abnormality, merging the labeling frames, and the like. The contour relation analysis comprises the steps of firstly, judging whether the contour length meets a certain threshold value through contour length calculation, and then, realizing detection through judging whether an outer contour exists; contour shape analysis is similar to contour relation analysis, except that after a contour length threshold value is met, whether a convex hull with a certain size exists in the contour is judged; the phenomenon that flaw marks are repeated often occurs in the abnormal mark summarization, and the overlapped mark frames are combined by adopting a principle of close accommodation, so that the phenomenon of repeated marking is avoided;
the image preprocessing is that the size of a convolution kernel of Gaussian filtering is 7 × 7, and the convolution kernel adopted by the image erosion operation of image secondary processing is 5 × 5;
FIG. 5 is a schematic diagram of the independent defect classification and extraction process, which shows the principle of the whole method and the corresponding implementation process in principle;
the following is a specific working process of the invention:
as shown in fig. 1, it is a flowchart of a defect feature extraction method, which is divided into three stages as a whole: the method comprises a model establishing stage, a chromaticity matching and extracting stage and an embedded convex hull analyzing and labeling stage; although the three stages are executed in series on the whole layout, the double-thread design is adopted during the embedded and convex hull analysis, and the parallel execution is realized, so that the program design framework accelerates the detection process of fabric flaws to a certain degree;
starting from model building, because model building is the basis for the implementation of the algorithm later in the method, in order to reduce the data amount processed by the algorithm, the input image is firstly subjected to down-sampling in the preprocessing process to reduce the size of the image by half, then the channels of the image are converted into an LAB space from RGB, and then Gaussian filtering with the convolution kernel size of 7 x 7 is carried out to achieve the preprocessing of the image; the model establishment is mainly realized by completing seed point selection, color block clustering and obtaining color block information, and corresponding color block clustering information can be obtained by correspondingly clustering each type of color block on a sample image, wherein the color block clustering information comprises information such as average LAB value of the color block, corresponding color block type number and the like;
the chroma matching and extracting process is carried out, the same model established before the chroma matching and extracting process is loaded, and the same preprocessing is carried out on the image to be detected in the process of loading the model, wherein the same preprocessing is required to be carried out on the image to be detected; after the preprocessing is finished, the preprocessed image pixels to be detected are traversed one by one and matched with model information, so that defects caused by chromaticity abnormality can be extracted, color blocks belonging to the model can be extracted respectively, and the embedded and convex hull analysis of the color block outline can be performed;
finally, independent images of various color blocks are obtained through the operations, and as shown in fig. 5, the position relation and the outline shape of the outline of each color block can be analyzed through an embedded and convex hull detection algorithm, so that a defect area containing abnormal relation and abnormal outline shape is obtained; after independent classification type feature extraction is carried out on fabric flaws, corresponding post-processing is carried out, and then the flaw detection process of the fabric is naturally achieved.
The invention realizes the purpose of detecting the fabric defects based on the independent classification type feature extraction technology, particularly aims at detecting the defects of fabrics containing discrete regular patterns or plain colors, and is believed to have wide application in practical working environments.
Claims (8)
1. A fabric flaw detection method based on independent classification type feature extraction is characterized by comprising the following steps:
step 1: preprocessing a sample fabric image by downsampling, channel conversion and Gaussian filtering;
step 2: carrying out artificial modeling on the fabric image preprocessed in the step 1 to obtain the color block type number and corresponding chromaticity information of the sample image;
and step 3: loading the model obtained in the step 2, processing the fabric image to be detected in the same preprocessing mode, marking abnormal areas through chromaticity matching and respectively extracting color blocks of various categories;
and 4, step 4: carrying out outline extraction on the color blocks extracted in the step 3 after image corrosion, and respectively sending the color blocks into embedded and convex hull detection algorithms to obtain and mark areas with abnormal outline relations and shapes;
and 5: and summarizing all types of defects, removing redundancy and combining the marking frames to obtain a defect detection result.
2. The method for detecting fabric defects based on independent classification type feature extraction according to claim 1, wherein the specific preprocessing details of the steps 1 and 4 are as follows: wherein the down-sampling scale factor is set to 0.5, the gamut in which the image processing is located is the LAB space, the gaussian filter convolution kernel is set to 7 x 7, and the convolution kernel for the erosion operation is set to 5 x 5.
3. The method for detecting fabric defects based on independent classification type feature extraction according to claim 1, wherein the specific steps of the step 2 are as follows:
step 2.1: selecting seed points on the sample image, specifically:
randomly selecting a pixel point on a certain color block of the sample image, and calculating the chromaticity distance colorDis value, which can be expressed as:
in the formula:
the respective accumulated mean values of the color blocks are calculated by
And (6) calculating.
Step 2.2: the method comprises the following steps: v is less than or equal to color DisthresholdTo determine whether two adjacent elements belong to the color block class, wherein vthresholdIs a distance threshold.
Step 2.3: by analogy, each type of color block on the sample image is clustered correspondingly, and corresponding color block clustering information, namely a model, can be obtained, wherein the color block clustering information comprises information such as an average LAB value of the color block and the corresponding color block type number.
4. The method for detecting fabric defects based on independent classification type feature extraction is characterized in that the distance threshold value setting of the step 2.2 is determined according to actual field conditions.
5. The method for detecting fabric defects based on independent classification type feature extraction according to claim 1, wherein the specific algorithm of the step 4 is as follows:
the image to be detected can be represented here as: i ═ I1+I2+I3+It+ N, wherein, I1Indicating a chroma abnormal part, I2Indicating a defect part of the inline type, I3Indicating a convex hull type defect part, ItRepresenting a normal texture part, and N representing a noise part;
wherein the chroma abnormal part I1Has been extracted in step 3 of claim 1;
for embedded type defect part I2And a convex hull type flaw part I3The two types of defect areas are extracted and marked through an embedded convex hull detection algorithm which analyzes the inclusion relation and the shape of the color block outline.
6. The method for detecting fabric defects based on independent classification type feature extraction as claimed in claim 1, wherein the merging of the marking frames in the step 5 is implemented by merging the overlapped marking frames by using a principle of close-tolerance, so as to avoid the phenomenon of repeated marking.
7. The near containment rule of claim 6 wherein the defective bins are merged above a threshold value to replace the original defective bins with the largest containing bins.
8. The method of claim 1, wherein the method is an independent classification type feature extraction method, and is more suitable for detecting defects of fabrics containing discrete regular patterns or plain colors.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111353364.4A CN113920112A (en) | 2021-11-18 | 2021-11-18 | Fabric flaw detection method based on independent classification type feature extraction |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111353364.4A CN113920112A (en) | 2021-11-18 | 2021-11-18 | Fabric flaw detection method based on independent classification type feature extraction |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113920112A true CN113920112A (en) | 2022-01-11 |
Family
ID=79246466
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111353364.4A Pending CN113920112A (en) | 2021-11-18 | 2021-11-18 | Fabric flaw detection method based on independent classification type feature extraction |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113920112A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117152158A (en) * | 2023-11-01 | 2023-12-01 | 海门市缔绣家用纺织品有限公司 | Textile defect detection method and system based on artificial intelligence |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109919899A (en) * | 2017-12-13 | 2019-06-21 | 香港纺织及成衣研发中心有限公司 | The method for evaluating quality of image based on multispectral imaging |
| CN110021023A (en) * | 2019-03-05 | 2019-07-16 | 西安工程大学 | A kind of electronics cloth defect segmentation method |
| CN111754470A (en) * | 2020-06-11 | 2020-10-09 | 厦门雨程户外运动用品有限公司 | Automatic cloth inspecting method and device, automatic cloth inspecting machine and storage medium |
| US20210343002A1 (en) * | 2020-07-28 | 2021-11-04 | Jiangnan University | Online Detection Method of Circular Weft Knitting Stripe Defects Based on Gray Gradient Method |
-
2021
- 2021-11-18 CN CN202111353364.4A patent/CN113920112A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109919899A (en) * | 2017-12-13 | 2019-06-21 | 香港纺织及成衣研发中心有限公司 | The method for evaluating quality of image based on multispectral imaging |
| CN110021023A (en) * | 2019-03-05 | 2019-07-16 | 西安工程大学 | A kind of electronics cloth defect segmentation method |
| CN111754470A (en) * | 2020-06-11 | 2020-10-09 | 厦门雨程户外运动用品有限公司 | Automatic cloth inspecting method and device, automatic cloth inspecting machine and storage medium |
| US20210343002A1 (en) * | 2020-07-28 | 2021-11-04 | Jiangnan University | Online Detection Method of Circular Weft Knitting Stripe Defects Based on Gray Gradient Method |
Non-Patent Citations (1)
| Title |
|---|
| 王永庆: "人工智能原理与方法 修订版", 31 August 2018, 西安交通大学出版社, pages: 381 - 382 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117152158A (en) * | 2023-11-01 | 2023-12-01 | 海门市缔绣家用纺织品有限公司 | Textile defect detection method and system based on artificial intelligence |
| CN117152158B (en) * | 2023-11-01 | 2024-02-13 | 海门市缔绣家用纺织品有限公司 | Textile defect detection method and system based on artificial intelligence |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zhu et al. | Yarn-dyed fabric defect detection based on autocorrelation function and GLCM | |
| CN114549522B (en) | Textile quality detection method based on target detection | |
| CN111179225B (en) | Test paper surface texture defect detection method based on gray gradient clustering | |
| CN111539935B (en) | Online cable surface defect detection method based on machine vision | |
| CN114820627B (en) | Textile quality detection method based on optical means | |
| CN113646801B (en) | Defect detection method, device and computer readable storage medium for defect image | |
| CN108562589A (en) | A method of magnetic circuit material surface defect is detected | |
| CN110889837A (en) | Cloth flaw detection method with flaw classification function | |
| CN112330628A (en) | Metal workpiece surface defect image detection method | |
| CN111047655A (en) | High-definition camera cloth defect detection method based on convolutional neural network | |
| CN113034488B (en) | Visual inspection method for ink-jet printed matter | |
| CN108181316A (en) | A kind of bamboo strip defect detection method based on machine vision | |
| CN115131353B (en) | Flat screen printing textile production abnormity identification and positioning method and system | |
| CN114494226B (en) | Method for detecting greasy dirt defect of spinning cake based on graph centroid tracking algorithm | |
| CN110781913A (en) | Zipper cloth belt defect detection method | |
| CN113920112A (en) | Fabric flaw detection method based on independent classification type feature extraction | |
| CN113989542A (en) | Texture image defect classification system and method based on multi-resolution feature fusion | |
| CN113936001B (en) | Textile surface flaw detection method based on image processing technology | |
| CN115266732A (en) | Carbon fiber tow defect detection method based on machine vision | |
| CN107256549A (en) | A kind of bamboo strip defect detection method based on machine vision | |
| CN113838040A (en) | Detection method for defect area of color texture fabric | |
| CN111402225A (en) | Cloth folding false detection defect discrimination method | |
| CN110322437B (en) | Fabric defect detection method based on automatic encoder and BP neural network | |
| CN113870299A (en) | 3D printing fault detection method based on edge detection and morphological image processing | |
| CN114202544A (en) | Complex workpiece defect detection method based on self-encoder |
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 |