CN111862027B - Textile flaw detection method based on low-rank sparse matrix decomposition - Google Patents
Textile flaw detection method based on low-rank sparse matrix decomposition Download PDFInfo
- Publication number
- CN111862027B CN111862027B CN202010678742.5A CN202010678742A CN111862027B CN 111862027 B CN111862027 B CN 111862027B CN 202010678742 A CN202010678742 A CN 202010678742A CN 111862027 B CN111862027 B CN 111862027B
- Authority
- CN
- China
- Prior art keywords
- matrix
- low
- flaw
- rank
- textile
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 239000011159 matrix material Substances 0.000 title claims abstract description 106
- 239000004753 textile Substances 0.000 title claims abstract description 83
- 238000001514 detection method Methods 0.000 title claims abstract description 67
- 238000000354 decomposition reaction Methods 0.000 title claims abstract description 59
- 230000007547 defect Effects 0.000 claims abstract description 26
- 238000000034 method Methods 0.000 claims abstract description 21
- 230000011218 segmentation Effects 0.000 claims abstract description 7
- 238000000638 solvent extraction Methods 0.000 claims abstract description 7
- 238000012549 training Methods 0.000 claims description 17
- 238000009826 distribution Methods 0.000 claims description 8
- 238000006243 chemical reaction Methods 0.000 claims description 6
- 238000010586 diagram Methods 0.000 claims description 5
- 230000003190 augmentative effect Effects 0.000 claims description 4
- 238000005457 optimization Methods 0.000 claims description 2
- 230000000737 periodic effect Effects 0.000 abstract description 3
- 239000004744 fabric Substances 0.000 description 8
- 238000004519 manufacturing process Methods 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 230000002950 deficient Effects 0.000 description 2
- 230000000877 morphologic effect Effects 0.000 description 2
- 238000007781 pre-processing Methods 0.000 description 2
- 230000003595 spectral effect Effects 0.000 description 2
- 238000007619 statistical method Methods 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000005311 autocorrelation function Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- 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
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
- G06V10/443—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components by matching or filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20021—Dividing image into blocks, subimages or windows
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20024—Filtering details
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—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)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Multimedia (AREA)
- Quality & Reliability (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
- Treatment Of Fiber Materials (AREA)
Abstract
The invention relates to the technical field of textile detection, in particular to a textile flaw detection method based on low-rank sparse matrix decomposition, which comprises the following steps of firstly partitioning a periodic textile according to periodic information to obtain flaw priori, guiding a low-rank decomposition model, and then adding Laplace regularization to increase the distance between a background and a defect area; finally, dividing the generated sparse matrix by adopting an optimal threshold segmentation algorithm to finish defect detection; the invention provides a textile detection method based on low-rank sparse matrix decomposition, which is characterized in that a flaw priori is determined through a block segmentation method, the flaw priori is added into a low-rank decomposition model, a low-rank decomposition model with weight is built, the detection precision of a large flaw block is increased, the distance between a flaw area and a background is increased through Laplace regularization terms, and the detection precision and robustness of flaws are further improved.
Description
Technical Field
The invention relates to the technical field of textile detection, in particular to a textile flaw detection method based on low-rank sparse matrix decomposition.
Background
P-textiles always develop various flaws during their production, and textile flaws are one of the main factors affecting the quality of textiles. Therefore, it is an indispensable step in the production of textiles to detect flaws. However, the research of the flaw detection algorithm is challenging due to the complex and changeable textures and various flaw types of the textile image.
At present, textiles can be largely divided into two categories: the first is a pure-color textile (such as plain weave and twill weave) with simple structure and no complex pattern; another type is a textile image having a periodically varying pattern, wherein one basic pattern is referred to as a block or a period, and different types of fabrics (e.g., box and star patterns) have different period sizes and shapes.
For the first type of textiles, there are many well established algorithms that are practically applicable, and they can be broadly divided into the following categories: 1) Statistical methods including autocorrelation functions, mathematical morphology, morphological filters, etc.; 2) Spectral methods, including Gabor filtering, wavelet transform, fourier transform; 3) The training method comprises the following steps: including neural networks, etc.; 4) Modeling methods include autoregressive models, markov random field models, and the like. Wherein, statistical methods and spectral methods have a greater false detection when detecting larger or smaller flaws.
The second type of textile still has a certain challenge in current detection due to the complexity of the texture unit. Recent detection algorithms include wavelet pre-processed golden image subtraction (WGIS), bragg Band (BB), regular Band (RB), ER algorithm, TC algorithm, LSG algorithm, and the like. While WGIS can detect large defects, it is ineffective for some small defects. For BB and RB, both the upper and lower bands of the Bollinger band are sensitive to any subtle changes in the input data and therefore are difficult to apply to practical industrial processes, e.g., over 70 types of defects in the fabric. ER is a sports spirit-based defect detection method, i.e. fair competition between different image blocks. TC divides the fabric into a block according to the periodic size of the fabric, then corrects the fabric image by using a block template, and finally extracts the texture characteristics of the block to detect defects. The LSG uses Morphological Component Analysis (MCA) to automatically segment the grid, and then converts the defect detection problem into a grid similarity problem, and although the LSG achieves a better effect, the adaptability of the LSG to different types of defects needs to be further improved.
The low-rank decomposition (LR) model can decompose the original image into a low-rank portion representing the image background and a sparse portion corresponding to the defect region. Patterned fabric images with complex texture elements have high visual redundancy and defective areas appear prominent in the fabric background. In view of these characteristics, it is more suitable to apply the low-rank decomposition model to textile flaw detection.
Disclosure of Invention
The invention aims to solve the technical problems that: in order to solve the problems that in the prior art, large uniform flaw blocks cannot be detected, and when the textile background and the flaw area are relatively consistent or the texture of a textile image is complex, the textile is difficult to separate by the traditional method, a textile detection method based on low-rank sparse matrix decomposition is provided, flaw priori is determined by a block segmentation method, the flaw priori is added into a low-rank decomposition model, a low-rank decomposition model with weight is built, the detection precision of the large flaw blocks is improved, the distance between the flaw area and the background is increased by a Laplace regularization term, and the detection precision and robustness of flaws are improved.
The technical scheme adopted for solving the technical problems is as follows: a textile flaw detection method based on low-rank sparse matrix decomposition comprises the following steps:
s1: inputting a flawless textile image containing a periodically varying pattern;
s2: determining the size of a pattern period template in the textile image, and partitioning the flawless textile image according to the size of the pattern period template to obtain a plurality of training feature blocks; specifically, the size of each feature block is the same as the size of the pattern period template;
s3: and extracting Gabor characteristics of each training characteristic block, calculating chebyshev distances among the training characteristic blocks, and constructing a characteristic distance matrix.
S4: calculating an average chebyshev distance d between training feature blocks 1 ；
S5: inputting a textile image to be detected, and partitioning the textile image to be detected according to the size of the pattern period template to obtain a plurality of detection feature blocks; extracting Gabor characteristics of each detection characteristic block, calculating chebyshev distances among the detection characteristic blocks, and constructing a characteristic distance matrix; calculating an average chebyshev distance d between the detected feature blocks 2 The method comprises the steps of carrying out a first treatment on the surface of the Wherein when d 2 ＞d 1 When the textile image to be detected is marked as a flaw block, otherwise, the textile image to be detected is marked as a flawless block, and the obtained result is flaw priori; the flaw prior is divided into a training stage and a testing stage, the steps S1-S4 are the training stage, the average chebyshev distance d1 between training characteristic blocks is obtained, the step S5 is the testing stage, the average chebyshev distance d2 between testing characteristic blocks is obtained, and the similarity between the characteristic blocks is represented by the distance between the characteristic blocks;
s6: extracting Gabor characteristics of each detection characteristic block in the textile image to be detected to form a characteristic matrix F;
s7: the flaw prior is subjected to dimension conversion and then is put into a low-rank decomposition and Laplacian regularization term model of the weight, and low-rank decomposition is carried out;
s8: and performing optimal threshold segmentation on the sparse matrix obtained by low-rank decomposition to obtain a final detection result.
Further, in step 6, the method of constructing the feature matrix F includes preprocessing the textile image to be detected with a Gabor filter to generate Gabor features, and then dividing the textile image to be detected into N detected feature blocks by using a Simple Linear Iterative Clustering (SLIC) algorithm, which is expressed as p= { P 1 ,P 2 ,…,P N Each detection feature block P i From the feature vector f i The representation, the textile image to be detected, can be represented as a feature matrix f= { F 1 ,f 2 ,…,f N }, whereinM is the dimension of the detection feature block, +.>Is the real number domain.
Further, in step 7, the method for constructing the weighted low-rank decomposition and Laplacian regularization term model is that the obtained flaw prior is put into an initial low-rank decomposition model through dimension conversion, the decomposition of the weighted low-rank decomposition model is guided, the detection rate of large homogeneous flaw blocks is increased, the detection precision of the model is improved, the weighted low-rank decomposition model is constructed,wherein W represents a flaw prior, F represents a characteristic matrix of the textile, B represents a low-rank matrix, represents a background area of an image of the textile to be detected, S represents a sparse matrix, and represents a flaw area.
When the difference between the background and the defect is not obvious or the texture of the textile image is complex, the correlation between the background and the defect is strong, so when the situation is faced, the detection precision of the method based on the low-rank decomposition model of the weight is not ideal, in order to solve the problem, further, in step 7, after the low-rank decomposition model of the weight is constructed, laplace regularization is introduced to enlarge the distance between the background and the defect so as to distinguish the defect area in the background, and the low-rank decomposition of the weight and the Laplace regularization term model are defined as follows:beta is the Lagrange multiplier and Θ is the Laplacian regularization.
In order to facilitate the solution, in step 7, after defining the low-rank decomposition and Laplacian regularization term model of the weights, an auxiliary variable H is introducedTo separate the objective function, the low rank decomposition of the weights and the Laplacian regularization term model are described asWherein W represents flaw prior, F represents a characteristic matrix of the textile, B represents a low-rank matrix, represents a background area of an image of the textile to be detected, S represents a sparse matrix, represents a flaw area, M F For the laplace matrix, tr () represents the trace operation of the matrix.
Further, in step S8, after decomposing the feature matrix F into the low-rank matrix B and the sparse matrix S, l of each column in the sparse matrix is obtained 1 The norm represents the significance of each detected feature block, and the greater the significance, the greater the probability of containing a flaw, thereby generating a flaw distribution map S; and denoising the flaw distribution diagram, converting the flaw distribution diagram into a gray image G, and finally dividing the gray image G by using an optimal threshold dividing algorithm to obtain a final binarization detection result.
The beneficial effects of the invention are as follows: according to the textile detection method based on low-rank sparse matrix decomposition, the defect priori is determined through the block segmentation method, the defect priori is added into the low-rank decomposition model, the low-rank decomposition model with weight is built, the detection precision of large defect blocks is increased, the distance between a defect area and a background is increased through Laplacian regular terms, the detection precision and the robustness of the defects are improved, and therefore the detection of the large defect blocks and the detection of textiles with relatively consistent background and defect areas or complex image textures can be achieved.
Drawings
The invention will be further described with reference to the drawings and examples.
FIG. 1 is a flow chart of the detection method of the present invention;
FIG. 2 is a flaw prior map of the present invention;
FIG. 3 is a low rank decomposition schematic of the present invention;
FIG. 4 is a schematic diagram of the detection result of the present invention;
Detailed Description
The invention is described in further detail below with reference to the accompanying drawings. It should be understood that these examples are illustrative of the present invention and are not intended to limit the scope of the present invention. Unless defined otherwise, technical or scientific terms used herein should be given the ordinary meaning as understood by one of ordinary skill in the art to which this invention belongs. Meanwhile, a detailed definition description of functions and constructions well known in the art will be omitted to make the present specification clearer and concise.
Example 1
As shown in fig. 1, a textile flaw detection method based on low-rank sparse matrix decomposition includes the following steps:
s1: inputting a flawless textile image containing a periodically varying pattern;
s2: determining the size of a pattern period template in the textile image, and partitioning the flawless textile image according to the size of the pattern period template to obtain a plurality of training feature blocks; specifically, the size of each feature block is the same as the size of the pattern period template;
s3: extracting Gabor characteristics of each training characteristic block, calculating chebyshev distances among the training characteristic blocks, and constructing a characteristic distance matrix;
in this step, gabor features of each training feature block are extracted using Gabor filters, and a Gabor filter bank defining a two-dimensional Gabor transform function is as follows:
wherein the parameter sigma x Sum sigma y Representing the shape factor of a gaussian surface; θ represents a direction; g 0 Representing the center frequency; the parameter θ is defined herein as 0, pi/4, pi/2 and 3pi/4; and x and y are coordinates.
S4: calculating an average chebyshev distance d between training feature blocks 1 ；
S5: inputting a textile image to be detected, and partitioning the textile image to be detected according to the size of the pattern period template to obtain a plurality of detection feature blocks; extracting Gabor characteristics of each detection characteristic block, calculating chebyshev distances among the detection characteristic blocks, and constructing a characteristic distance matrix;
in this step, gabor features of each detection feature block are extracted using Gabor filters, and a Gabor filter bank defining a two-dimensional Gabor transform function is as follows:
calculating an average chebyshev distance d between the detected feature blocks 2 ；
Wherein when d 2 ＞d 1 The textile image to be detected is marked as a flaw block, otherwise, the textile image to be detected is marked as a flawless block, and the flaw prior is obtained as a flaw prior as shown in fig. 2;
s6: extracting Gabor characteristics of each detection characteristic block in the textile image to be detected to form a characteristic matrix F;
preprocessing a textile image to be detected by using a Gabor filter to generate Gabor characteristics, and then dividing the textile image to be detected into N detection characteristic blocks by using a Simple Linear Iterative Clustering (SLIC) algorithm, wherein the N detection characteristic blocks are expressed as P= { P 1 ,P 2 ,…,P N Each detection feature block P i From the feature vector f i The representation, the textile image to be detected, can be represented as a feature matrix f= { F 1 ,f 2 ,…,f N }, whereinM is the dimension of the detection feature block, +.>Is the real number domain;
s7: the flaw prior is subjected to dimension conversion and then is put into a low-rank decomposition and Laplacian regularization term model of the weight, and low-rank decomposition is carried out, wherein the decomposition process is shown in figure 3;
the method comprises the following steps:
the initial low rank decomposition model is:
wherein F represents a characteristic matrix of the textile, B represents a low-rank matrix, S represents a sparse matrix, S represents a flaw area, and lambda is a Lagrange multiplier;
in order to increase the detection precision of large flaw blocks and improve the detection precision of a model, flaw priori is added, and a low-rank decomposition model with weight is constructed as follows:
wherein W represents a flaw prior, F represents a characteristic matrix of the textile, B represents a low-rank matrix, S represents a sparse matrix, S represents a flaw region, and lambda is a Lagrangian multiplier;
the augmented lagrangian function is:
wherein μ is a lagrange multiplier, W represents a flaw prior, F represents a characteristic matrix of the textile, B represents a low-rank matrix, S represents a background area of the textile image to be detected, S represents a sparse matrix, λ is a lagrange multiplier, tr represents a trace of the matrix, and is a norm, and is a pi 1 The number of norms of 1 is indicated, I F Representing the F norm;
when the difference between the background and the defect is not obvious or the texture of the fabric image is complex, the correlation between the background and the defect is strong. Therefore, in the face of this, the detection accuracy of the method of the weight-based low-rank decomposition model is not ideal. To solve this problem, laplace regularization was introduced to expand the distance between the background and the defect in order to distinguish between defective areas in the background, defined as follows:
wherein N represents the number of regional blocks, beta is Lagrange multiplier, theta is Laplacian regularization operator, S represents sparse matrix, represents flaw region, S i Represents an i-th block sparse matrix, a i,j Representing elements of the affinity matrix A, M F For a Laplace matrix, tr represents the trace of the matrix;
the affinity matrix a is defined as follows:
wherein a is i,j Elements representing an affinity matrix A, I i Representing the i-th block image area, f i Representing the average gray level of the i-th block image area, sigma representing the variance of the whole image, and V representing the set of adjacent block pairs on the image;
laplacian matrix M F The (i, j) th item is defined as:
on the basis of a low-rank decomposition model of the weight, laplace regularization is added to construct a low-rank decomposition and Laplace regularization term model of the weight, and the model is defined as follows:
wherein beta is Lagrange multiplier, Θ is Laplacian regularization operator, W represents flaw priori, F represents characteristic matrix of textile, B represents low-rank matrix, represents background area of textile image to be detected, S represents sparse matrix, and represents flaw area;
for ease of solution, the introduction of the auxiliary variable H to separate the objective function, the low-rank decomposition of the weights and the laplace regularization term modulus can be described as:
s.t.F＝B+S,S＝H
wherein beta is Lagrangian multiplier, B is low-rank matrix, represents background area of textile image to be detected, W is flaw prior, F is characteristic matrix of textile, S is sparse matrix, represents flaw area, M F For the Laplace matrix, tr () represents the trace operation of the matrix;
currently, there are several mainstream algorithms for solving low-rank decomposition, such as accelerating near-end gradient (APG), augmented Lagrangian Multiplier (ALM), and Alternating Direction (ADM). Considering the efficiency and precision of the algorithm, the low-rank sparse matrix decomposition model of the convex optimization model is solved by adopting an ADM method, and the minimum augmented Lagrangian function is as follows:
wherein Y is 1 And Y 2 The matrix is Lagrange multiplier matrix, lambda, mu, beta is Lagrange multiplier, B represents low-rank matrix, represents background area of textile image to be detected, W represents flaw priori, F represents characteristic matrix of textile, S represents sparse matrix, represents flaw area, M F The matrix is Laplace matrix, tr represents the trace of the matrix, and H is an auxiliary variable matrix;
since the above problem is unconstrained, other variables (e.g., S, H) can be fixed, iteratively searching for the optimal values of B, S, H, respectively, and then updating the lagrangian multiplier;
s8: performing optimal threshold segmentation on a sparse matrix obtained by low-rank decomposition to obtain a final detection result;
after decomposing the feature matrix F into a low-rank matrix B and a sparse matrix S, solving l of each column in the sparse matrix 1 Norms to represent the saliency of each detected feature block:
Sal(I i )＝||S i || 1
wherein S is i Represents the I-th sparse matrix, I i Representing a saliency map corresponding to the ith sparse matrix;
if S i || 1 The larger the i-th sparse matrix is, the most likely contains flaws, and the new flaw distribution can be obtained by denoising the flaw distribution map generated by the sparse matrix S
Wherein g is a convolution template, x represents convolution operation, and o represents dot product operation;
will produce a defect profileConversion into a gray-scale map G:
and finally, dividing the G by using an optimal threshold dividing algorithm to obtain a final binarization detection result, wherein the detection result is shown in fig. 4.
The above-described preferred embodiments according to the present invention are intended to suggest that, from the above description, various changes and modifications can be made by the worker in question without departing from the technical spirit of the present invention. The technical scope of the present invention is not limited to the description, but must be determined according to the scope of claims.
Claims (4)
1. A textile flaw detection method based on low-rank sparse matrix decomposition is characterized by comprising the following steps of: the method comprises the following steps:
s1: inputting a flawless textile image containing a periodically varying pattern;
s2: determining the size of a pattern period template in the textile image, and partitioning the flawless textile image according to the size of the pattern period template to obtain a plurality of training feature blocks;
s3: extracting Gabor characteristics of each training characteristic block, calculating chebyshev distances among the training characteristic blocks, and constructing a characteristic distance matrix;
s4: calculating an average chebyshev distance d between training feature blocks 1 ；
S5: inputting a textile image to be detected, and partitioning the textile image to be detected according to the size of the pattern period template to obtain a plurality of detection feature blocks; extracting Gabor characteristics of each detection characteristic block, calculating chebyshev distances among the detection characteristic blocks, and constructing a characteristic distance matrix; calculating an average chebyshev distance d between the detected feature blocks 2 The method comprises the steps of carrying out a first treatment on the surface of the Wherein when d 2 ＞d 1 When the textile image to be detected is marked as a flaw block, otherwise, the textile image to be detected is marked as a flawless block, and the obtained result is flaw priori;
s6: extracting Gabor characteristics of each detection characteristic block in the textile image to be detected to form a characteristic matrix F;
s7: the flaw prior is subjected to dimension conversion and then is put into a low-rank decomposition and Laplacian regularization term model of the weight, and low-rank decomposition is carried out;
in step 7, the method for constructing the weighted low-rank decomposition and Laplacian regularization term model is that the obtained flaw prior is put into an initial low-rank decomposition model through dimension conversion, the weighted low-rank decomposition model is constructed as follows,wherein B represents a low-rank matrix, represents a background area of a textile image to be detected, W represents a flaw prior, S represents a sparse matrix, represents a flaw area, F represents a feature matrix of the textile image to be detected, and lambda is a Lagrange multiplier;
in step 7, after defining the low-rank decomposition and Laplace regularization term model of the weights, an auxiliary variable H is introduced to separate the objective function, and the low-rank decomposition and Laplace regularization term model of the weights is described asWherein M is F For the Laplace matrix, tr () represents the trace operation of the matrix;
adopting an ADM method to solve a convex optimization model low-rank sparse matrix decomposition model, wherein the minimum augmented Lagrange function is as follows:
wherein Y is 1 And Y 2 The matrix is Lagrange multiplier matrix, lambda, mu, beta is Lagrange multiplier, B represents low-rank matrix, represents background area of textile image to be detected, W represents flaw priori, F represents characteristic matrix of textile, S represents sparse matrix, represents flaw area, M F The matrix is Laplace matrix, tr represents the trace of the matrix, and H is an auxiliary variable matrix;
s8: and performing optimal threshold segmentation on the sparse matrix obtained by low-rank decomposition to obtain a final detection result.
2. The textile flaw detection method based on low-rank sparse matrix decomposition according to claim 1, wherein: in step 6, the method for constructing the feature matrix F is that the textile image to be detected is preprocessed by using a Gabor filter to generate Gabor features, and then the textile to be detected is clustered by using a simple linear iterative clustering SLIC algorithmThe image is partitioned into N detected feature blocks, denoted as p= { P 1 ,P 2 ,…,P N Each detection feature block P i From the feature vector f i Representing, namely representing the textile image to be detected as a feature matrix F= { F 1 ,f 2 ,…,f N F.epsilon.R }, where F.epsilon.R M×N M is the dimension of the detection feature block and R is the real number domain.
3. The textile flaw detection method based on low-rank sparse matrix decomposition according to claim 1, wherein: in step 7, after constructing the low-rank decomposition model of the weight, laplace regularization is introduced to enlarge the distance between the background and the defect, and the low-rank decomposition and Laplace regularization term model of the weight is defined as follows:beta is the Lagrange multiplier and Θ is the Laplacian regularization.
4. A textile flaw detection method based on low rank sparse matrix decomposition according to any of claims 1-3, wherein: in step S8, after decomposing the feature matrix F into a low-rank matrix B and a sparse matrix S, calculating l of each column in the sparse matrix 1 The norm represents the significance of each detected feature block, and the greater the significance, the greater the probability of containing a flaw, thereby generating a flaw distribution map S; and denoising the flaw distribution diagram, converting the flaw distribution diagram into a gray image G, and finally dividing the gray image G by using an optimal threshold dividing algorithm to obtain a final binarization detection result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010678742.5A CN111862027B (en) | 2020-07-15 | 2020-07-15 | Textile flaw detection method based on low-rank sparse matrix decomposition |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010678742.5A CN111862027B (en) | 2020-07-15 | 2020-07-15 | Textile flaw detection method based on low-rank sparse matrix decomposition |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111862027A CN111862027A (en) | 2020-10-30 |
| CN111862027B true CN111862027B (en) | 2024-02-20 |
Family
ID=72983458
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010678742.5A Active CN111862027B (en) | 2020-07-15 | 2020-07-15 | Textile flaw detection method based on low-rank sparse matrix decomposition |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111862027B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112308882B (en) * | 2020-11-10 | 2024-07-02 | 苏州兴钊防务研究院有限公司 | Video moving object detection method and device based on structural decomposition |
| CN113033318B (en) * | 2021-03-01 | 2023-09-26 | 深圳大学 | Human body motion detection method, device and computer readable storage medium |
| CN116109640B (en) * | 2023-04-13 | 2023-08-18 | 常州微亿智造科技有限公司 | Workpiece surface small defect detection method in industrial detection |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR101556603B1 (en) * | 2014-12-30 | 2015-10-01 | 연세대학교 산학협력단 | Apparatus and Method for Image Seperation using Rank Prior |
| CN107358245A (en) * | 2017-07-19 | 2017-11-17 | 安徽大学 | Method for detecting image collaborative salient region |
| CN107392921A (en) * | 2017-07-14 | 2017-11-24 | 西安邮电大学 | A kind of semi-supervised multi-object clustering image partition method based on Chebyshev's distance |
| CN107705306A (en) * | 2017-10-26 | 2018-02-16 | 中原工学院 | A kind of fabric defect detection method based on multiple features matrix low rank decomposition |
| CN109712113A (en) * | 2018-11-28 | 2019-05-03 | 中原工学院 | A kind of fabric defect detection method based on cascade low-rank decomposition |
| CN110188770A (en) * | 2019-05-17 | 2019-08-30 | 重庆邮电大学 | A kind of non-convex low-rank well-marked target detection method decomposed based on structure matrix |
| CN111046868A (en) * | 2019-11-26 | 2020-04-21 | 广东石油化工学院 | Target significance detection method based on matrix low-rank sparse decomposition |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108324263B (en) * | 2018-01-11 | 2020-05-08 | 浙江大学 | Noninvasive cardiac electrophysiology inversion method based on low-rank sparse constraint |
-
2020
- 2020-07-15 CN CN202010678742.5A patent/CN111862027B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR101556603B1 (en) * | 2014-12-30 | 2015-10-01 | 연세대학교 산학협력단 | Apparatus and Method for Image Seperation using Rank Prior |
| CN107392921A (en) * | 2017-07-14 | 2017-11-24 | 西安邮电大学 | A kind of semi-supervised multi-object clustering image partition method based on Chebyshev's distance |
| CN107358245A (en) * | 2017-07-19 | 2017-11-17 | 安徽大学 | Method for detecting image collaborative salient region |
| CN107705306A (en) * | 2017-10-26 | 2018-02-16 | 中原工学院 | A kind of fabric defect detection method based on multiple features matrix low rank decomposition |
| CN109712113A (en) * | 2018-11-28 | 2019-05-03 | 中原工学院 | A kind of fabric defect detection method based on cascade low-rank decomposition |
| CN110188770A (en) * | 2019-05-17 | 2019-08-30 | 重庆邮电大学 | A kind of non-convex low-rank well-marked target detection method decomposed based on structure matrix |
| CN111046868A (en) * | 2019-11-26 | 2020-04-21 | 广东石油化工学院 | Target significance detection method based on matrix low-rank sparse decomposition |
Non-Patent Citations (6)
| Title |
|---|
| Fabric defect inspection using prior knowledge guided least squares regression;Junjie Cao;Multimed Tools Appl;正文第4141-4155页 * |
| 基于低秩分解的织物疵点检测;杨恩君;纺织学报;第41卷(第5期);全文 * |
| 基于低秩稀疏矩阵分解的织物疵点检测算法研;高广帅;中国优秀硕士学位论文全文数据库工程科技Ⅰ辑(第7期);正文第59页 * |
| 基于局部最优分析的纺织品瑕疵检测方法;刘威;常兴治;梁久祯;贾靓;顾程熙;;模式识别与人工智能;20180215(第02期);全文 * |
| 基于模板校正与低秩分解的纺织品瑕疵检测方法;纪旋;梁久祯;侯振杰;常兴治;刘威;;模式识别与人工智能;20190315(第03期);全文 * |
| 基于迹表示和正则化的显著目标检测算法;马晓迪;计算机辅助设计与图形学学报;第30卷(第11期);正文第2018-2025页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111862027A (en) | 2020-10-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111862027B (en) | Textile flaw detection method based on low-rank sparse matrix decomposition | |
| Hu | Automated defect detection in textured surfaces using optimal elliptical Gabor filters | |
| Zheng et al. | A generic semi-supervised deep learning-based approach for automated surface inspection | |
| CN105678788B (en) | A kind of fabric defect detection method based on HOG and low-rank decomposition | |
| Bu et al. | Detection of fabric defects by auto-regressive spectral analysis and support vector data description | |
| CN111724372A (en) | Method, terminal and storage medium for detecting cloth defects based on antagonistic neural network | |
| CN110245593A (en) | A kind of images of gestures extraction method of key frame based on image similarity | |
| CN110310277B (en) | Fabric defect detection method based on depth feature and NTV-RPCA | |
| CN107274385A (en) | Textile flaw detection method based on class grid pattern and its Gabor characteristic | |
| Sadaghiyanfam | Using gray-level-co-occurrence matrix and wavelet transform for textural fabric defect detection: A comparison study | |
| Liu et al. | A classification method of glass defect based on multiresolution and information fusion | |
| Lin et al. | Defect enhancement generative adversarial network for enlarging data set of microcrack defect | |
| CN111091134A (en) | Method for identifying tissue structure of colored woven fabric based on multi-feature fusion | |
| Liang et al. | Automatic defect detection of texture surface with an efficient texture removal network | |
| CN105184777A (en) | Painted design fabric defect detection method based on image decomposition | |
| Sulistianingsih et al. | Classification of batik image using grey level co-occurrence matrix feature extraction and correlation based feature selection | |
| CN115147341A (en) | Fabric surface defect classification depth network method based on biological vision inspiration | |
| Jia et al. | Fabric defect inspection based on lattice segmentation and lattice templates | |
| Oni et al. | Patterned fabric defect detection and classification (FDDC) techniques: a review | |
| Dong et al. | Fusing multilevel deep features for fabric defect detection based NTV-RPCA | |
| Bao et al. | Low-rank decomposition fabric defect detection based on prior and total variation regularization | |
| CN116842210B (en) | Textile printing texture intelligent retrieval method based on texture features | |
| CN113838040A (en) | Detection method for defect area of color texture fabric | |
| Yu et al. | Research on CNN Algorithm for Monochromatic Fabric Defect Detection | |
| Maragos et al. | Image analysis of soil micromorphology: feature extraction, segmentation, and quality inference |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |