CN114494259A - Cloth defect detection method based on artificial intelligence - Google Patents
Cloth defect detection method based on artificial intelligence Download PDFInfo
- Publication number
- CN114494259A CN114494259A CN202210401115.6A CN202210401115A CN114494259A CN 114494259 A CN114494259 A CN 114494259A CN 202210401115 A CN202210401115 A CN 202210401115A CN 114494259 A CN114494259 A CN 114494259A
- Authority
- CN
- China
- Prior art keywords
- cloth
- gray
- image
- complexity
- scale image
- 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.)
- Granted
Links
- 239000004744 fabric Substances 0.000 title claims abstract description 163
- 230000007547 defect Effects 0.000 title claims abstract description 40
- 238000013473 artificial intelligence Methods 0.000 title claims abstract description 19
- 238000001514 detection method Methods 0.000 title claims abstract description 17
- 230000000737 periodic effect Effects 0.000 claims abstract description 60
- 230000002950 deficient Effects 0.000 claims abstract description 6
- 238000000034 method Methods 0.000 claims description 24
- 239000011159 matrix material Substances 0.000 claims description 22
- 238000001228 spectrum Methods 0.000 claims description 8
- 239000013598 vector Substances 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 241001270131 Agaricus moelleri Species 0.000 description 1
- 238000012935 Averaging Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000004907 flux Effects 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 230000011218 segmentation 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
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/40—Analysis of texture
- G06T7/41—Analysis of texture based on statistical description of texture
- G06T7/45—Analysis of texture based on statistical description of texture using co-occurrence matrix computation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
-
- 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)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Computing Systems (AREA)
- Quality & Reliability (AREA)
- Mathematical Physics (AREA)
- Probability & Statistics with Applications (AREA)
- Geometry (AREA)
- Image Analysis (AREA)
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
- Image Processing (AREA)
Abstract
The invention relates to the technical field of artificial intelligence, in particular to a cloth defect detection method based on artificial intelligence, which comprises the following steps: dividing the gray level image into different grades according to the gray level range of the gray level image of the cloth to obtain the gray level image of the cloth with different grades; carrying out region division on the cloth gray-scale image to obtain a plurality of periodic regions; calculating the sum of the texture complexity of all periodic regions on the cloth gray-scale image to obtain the texture complexity of the cloth gray-scale image; selecting a piece of cloth gray-scale image with the optimal grade according to the texture complexity of the piece of cloth gray-scale image, and marking the piece of cloth gray-scale image as an optimal piece of cloth image; and calculating the defect rate of the cloth according to the optimal cloth image, wherein the cloth with the defect rate larger than a set threshold value is defective. The invention can improve the accuracy of cloth defect detection.
Description
Technical Field
The invention relates to the technical field of artificial intelligence, in particular to a cloth defect detection method based on artificial intelligence.
Background
Cloth is an essential article in our lives, and due to wide application range, people pay great attention to the quality of the cloth. Cloth manufacturing enterprises mainly stand in front of cloth inspecting equipment through professional cloth inspectors, find cloth cover defects through naked eyes and then mark or record the defects, and therefore a large amount of manpower is consumed. In addition, a method for acquiring the texture of the cloth by utilizing the gray level co-occurrence matrix and determining the cloth defect by comparing the texture of the cloth is also provided.
However, if this method is applied, gray scale grading is required when generating the gray co-occurrence matrix, and the grading is linearly divided based on a given grading number, and different textures obtain different gray scales in the cloth image under different illumination, and the grading mode based on the given grade number cannot adapt to all textures: too little grading can result in some gray level co-occurrence matrices not accurately representing some complex textures; and too much grading can cause too much calculation of the gray level co-occurrence matrix and difficulty in extracting texture features of the cloth.
Disclosure of Invention
In order to solve the technical problems, the invention aims to provide a cloth defect detection method based on artificial intelligence, and the adopted technical scheme is as follows:
dividing the gray level image into different grades according to the gray level range of the gray level image of the cloth to obtain the gray level image of the cloth with different grades;
acquiring frequency spectrum information of any piece of cloth gray-scale image, and performing region division on the piece of cloth gray-scale image according to the frequency spectrum information to obtain a plurality of periodic regions;
obtaining the gradient value and coordinates of each pixel point in a period region to form label data of the period region, and classifying the label data to obtain a plurality of classes, wherein each class corresponds to a local region;
acquiring the edge complexity and the internal complexity of each local area, and acquiring the texture complexity of the periodic area according to the edge complexity, the internal complexity and the area of all the local areas;
calculating the sum of the texture complexity of all periodic regions on the cloth gray-scale image to obtain the texture complexity of the cloth gray-scale image, and acquiring the maximum texture complexity in all levels of the cloth gray-scale image;
obtaining the optimal degree of the cloth gray-scale image at the current level according to the difference between the areas of the periodic regions in the cloth gray-scale image at the current level and the cloth gray-scale image at the adjacent level and the difference between the texture complexity of the cloth gray-scale image at the current level and the maximum texture complexity;
acquiring a cloth gray scale image of a grade corresponding to the maximum value of the optimal degree, and recording the image as an optimal cloth image; and calculating the defect rate of the cloth according to the optimal cloth image, wherein the cloth with the defect rate larger than a set threshold value is defective.
Preferably, the area division of the cloth gray-scale image is specifically as follows:
acquiring frequency spectrum information of any piece of cloth gray-scale image to obtain a gray-scale amplitude-frequency image, and establishing an amplitude-frequency coordinate system on the gray-scale amplitude-frequency image by taking the central point of the image as an original point, the transverse direction as an x axis and the longitudinal direction as a y axis; respectively selecting maximum points on an x axis and a y axis as a transverse periodic point and a longitudinal periodic point; obtaining the width and the length according to the reciprocal of the frequency of the transverse periodic points and the longitudinal periodic points; and dividing the cloth gray scale image into a plurality of periodic areas according to the width and the length.
Preferably, the dividing of different levels according to the gray value range of the cloth gray image is specifically: and all 256 divisible integers in the gray value range of the cloth gray image are used as division levels, and the cloth gray image is divided into different levels.
Preferably, the method for acquiring the edge complexity of the local area specifically includes:
wherein A isLThe edge complexity of the local area, N is the number of the edge pixel points of the local area; calculating the hessian matrix of each edge pixel point in the local area,
representing the eigenvector corresponding to the maximum eigenvalue of the Hessian matrix of the nth edge pixel point,
hessian moment representing n-1 th edge pixel pointThe eigenvector corresponding to the largest eigenvalue of the array,
to represent
And
similarity of two feature vectors.
Preferably, the method for acquiring the internal complexity of the local region specifically includes:
wherein A isOIs the internal complexity of the local region, M is the number of pixels inside the local region,
representing the gradient value of the mth pixel point in the local area,
representing the gradient value of the m-1 th pixel point in the local area,
representing the similarity of the gradient values of the two pixels.
Preferably, the method for acquiring the texture complexity of the periodic region specifically includes:
wherein A is the texture complexity of the periodic region,
indicating the edge complexity of the pth local area,
represents the internal complexity, S, of the p-th regionpThe area of the p-th local region is shown.
Preferably, the method for acquiring the preference degree specifically includes:
wherein, PiRepresenting the degree of preference of a gray-scale image of the cloth of level i, ai、biRespectively representing the width and length of a periodic region in a gray-scale image of a piece of cloth of level ii+1、bi+1Respectively representing the width and length of a period region on a gray-scale image of the cloth with the level of i +1, (a)i*bi)－(ai+1*bi+1) Indicates the difference between the areas of the periodic regions in the gray-scale image of the piece at the level i and the gray-scale image of the piece at the level i +1, AFiTexture complexity, A, for a piece goods gray-scale image of level iFMRepresenting the maximum texture complexity.
Preferably, the method for acquiring the defect rate of the cloth specifically comprises the following steps:
acquiring a gray level co-occurrence matrix of each period region in the optimal cloth image, calculating an entropy value of the gray level co-occurrence matrix of each period region, D (ENT) representing the variance of entropy values of all period regions, and E (ENT) representing the mean value of entropy values of all navicular regions.
The embodiment of the invention at least has the following beneficial effects:
according to the method, the texture complexity of the cloth gray-scale image is obtained by calculating the edge complexity and the internal complexity of the local area in the cloth gray-scale image, the optimization degree of the image is further determined according to the texture complexity, the cloth gray-scale image with the proper grading grade is selected to obtain the gray-scale co-occurrence matrix, and finally the defects of the cloth are detected. The method can accurately represent the texture of the current cloth, cannot cause overlarge calculated amount of the gray level co-occurrence matrix to be difficult to extract the texture features of the cloth due to excessive grading quantity, and simultaneously improves the accuracy of cloth defect detection.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions and advantages of the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.
FIG. 1 is a flow chart of a method for detecting defects in cloth based on artificial intelligence.
Detailed Description
To further illustrate the technical means and effects of the present invention adopted to achieve the predetermined objects, the following detailed description will be given to a cloth defect detection method based on artificial intelligence according to the present invention, and its specific implementation, structure, features and effects thereof, with reference to the accompanying drawings and preferred embodiments. In the following description, the different references to "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, the particular features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
The following describes a specific scheme of the cloth defect detection method based on artificial intelligence in detail with reference to the accompanying drawings.
Example 1:
the invention aims at the following scenes: the device is used for detecting the defects of the cloth with various weaves on a cloth inspecting machine.
Referring to fig. 1, a flowchart of steps of a method for detecting defects of a piece of cloth based on artificial intelligence according to an embodiment of the present invention is shown, where the method includes the following steps:
firstly, dividing the gray level image into different levels according to the gray level range of the gray level image of the cloth to obtain the gray level image of the cloth with different levels.
Specifically, a camera is used on the cloth inspection machine to acquire a grey scale image of the cloth. It should be noted that the camera is preferably a grayscale camera because the grayscale camera collects photons of all wavelengths, while the color camera only receives photons of three RGB bands and performs a demosaicing neighborhood averaging operation, so that the light flux and detail performance are weaker than those of the grayscale camera.
The set gray-scale value range is denoted as gradation range [ G1,Gk]Wherein G is1Not less than the minimum value of pixel values, G, on the grey scale image of the clothkAnd cannot be larger than the maximum value of the pixel values on the cloth gray image. All 256 divisible integers in the range of the hierarchy
As a grading, dividing the cloth gray image into
And (5) grading to obtain cloth gray scale images under different grades. For example, a specific method for dividing a cloth gray image into 8 levels is as follows: dividing the pixel value of each pixel point in the cloth gray image by 32 and then rounding to obtain a value serving as a new pixel value to generate a new image, wherein the new image is the cloth gray image with the grade of 8.
And then, acquiring the frequency spectrum information of any piece of cloth gray-scale image, and performing region division on the piece of cloth gray-scale image according to the frequency spectrum information to obtain a plurality of periodic regions. The periodic region is a region in which the pattern of the cloth appears to change periodically on the gray-scale image of the cloth. In the same periodic region, the pattern does not repeat.
Specifically, a Fourier transform is used for the cloth gray-scale image to obtain a gray-scale amplitude-frequency image. On the gray scale amplitude frequency image, toThe central point of the image is the origin, the transverse direction is the x axis, the longitudinal direction is the y axis, and an amplitude-frequency coordinate system is established. Respectively counting maximum value points on an x axis and a y axis as points to be selected; selecting the largest point of all points to be selected on the x axis as a transverse period point; and selecting the largest one of the points to be selected on the y axis as the longitudinal periodic point. Acquiring the frequency f in the corresponding gray scale amplitude-frequency diagram of the transverse periodic pointtAcquiring the frequency f in the gray scale amplitude-frequency diagram corresponding to the longitudinal periodic pointl. The size of the periodic region in the cloth gray-scale image is as follows:
wherein a is the width of the periodic region and b is the length of the periodic region. Dividing the cloth gray-scale image into different periodic regions by taking a as the width and b as the length.
It should be noted that this embodiment shows only one embodiment of dividing the area of the cloth gray-scale image, and the implementer may select other methods such as semantic segmentation according to the actual situation.
Then, obtaining the label data of a periodic area formed by the gradient value and the coordinates of each pixel point in the periodic area, and classifying the label data to obtain a plurality of classes, wherein each class corresponds to a local area; and acquiring the edge complexity and the internal complexity of each local area, and acquiring the texture complexity of the periodic area according to the edge complexity, the internal complexity and the area of all the local areas.
Specifically, for a period region on the cloth gray-scale image, coordinates (x, y) of each pixel point in the period region are obtained, the gradient value ∇ G of each pixel point is calculated, and description D of each pixel point is generatedi=(x,y,∇Gi) The descriptions of all the pixels in the periodic region together form label data D = (D) of the periodic region1,D2,…,Di…). In the present embodiment, it is preferred that,and processing the label data by using mean shift clustering, and dividing pixel points in the periodic region into different category clusters, wherein each category cluster corresponds to a local region.
The method for acquiring the edge complexity of the local area specifically comprises the following steps:
the edge complexity refers to the complexity of an edge pixel point in a certain local area in a periodic area on a cloth gray scale image. For each local area in each period area, acquiring edge pixel points of the local area, and calculating the edge complexity of the local area:
wherein A isLThe edge complexity of the local area, N is the number of the edge pixel points of the local area; calculating the hessian matrix of each edge pixel point in the local area,
representing the eigenvector corresponding to the maximum eigenvalue of the Hessian matrix of the nth edge pixel point,
representing the eigenvector corresponding to the maximum eigenvalue of the Hessian matrix of the n-1 th edge pixel point,
to represent
And
similarity of the two feature vectors. In the present embodiment, it is preferred that,
expressed is cosine similarity, expressing the curvature difference between two adjacent pixel points, wherein the smaller the difference is, the greater the similarity is; this is achieved byThe greater the difference in the directions of maximum curvature of adjacent edge points, the greater the edge complexity in the local area.
The method for acquiring the internal complexity of the local area specifically comprises the following steps:
wherein A isOIs the internal complexity of the local region, M is the number of pixels inside the local region,
representing the gradient value of the mth pixel point in the local area,
representing the gradient value of the m-1 th pixel point in the local area,
representing the similarity of the gradient values of the two pixels. In the present embodiment, it is preferred that,
the cosine similarity represents the difference of two adjacent pixel points on the cloth gray-scale image, and the smaller the difference is, the larger the value is. Inside the local area, the greater the difference in gray scale gradients, the greater the complexity within the local area.
Calculating the texture complexity of the periodic region:
wherein A is the texture complexity of the periodic region,
indicating the edge complexity of the pth local area,
represents the internal complexity, S, of the p-th regionpThe area of the p-th local region is shown. In the same periodic region, the larger the number of the local regions, the greater the edge complexity, the greater the local complexity, the smaller the area of the local region, and the greater the texture complexity of the periodic region.
Further, calculating the sum of the texture complexity of all periodic areas on the cloth gray-scale image to obtain the texture complexity of the cloth gray-scale image, and acquiring the maximum texture complexity in all levels of the cloth gray-scale image; and obtaining the optimal degree of the cloth gray-scale image at the current level according to the difference between the areas of the periodic regions in the cloth gray-scale image at the current level and the cloth gray-scale image at the adjacent level and the difference between the texture complexity of the cloth gray-scale image at the current level and the maximum texture complexity.
Specifically, in a cloth gray scale image at the same level, the edge complexity and the internal complexity of all local areas in each period area are obtained, and the texture complexity of all period areas is calculated; the sequence number q of the periodic region, the texture complexity A of the periodic region and the description R of the periodic region are formedq= (q, A), the descriptions of all period areas in the cloth gray scale image at the same level together form a sample set R = (R)1,R2,…,Rq…) fitting the positions of all periodic regions in the same gray-scale image in a sample space into a straight line by using RANSAC fitting; obtaining the projection value A of the straight line in the sample space on the texture complexityFI.e. the projection value is the texture complexity A of the piece gray level imageFAnd obtaining the maximum value of the texture complexity of the cloth gray-scale images of all levels, and recording the maximum value as the maximum texture complexity AFM。
It should be noted that, if a defect exists in the grayscale image, the texture complexity of the periodic region where the defect is located is definitely different from the texture complexity of other periodic regions, that is, the texture complexity may be affected by the defect to cause an error, and the method adopted in this embodiment may enable the texture complexity of the finally obtained grayscale image to be unaffected by the local defect.
Another embodiment of obtaining the texture complexity of the cloth gray scale image is as follows: and calculating the sum of the texture complexity of all periodic areas on the cloth gray-scale image to obtain the texture complexity of the cloth gray-scale image.
The calculation method of the preference degree specifically comprises the following steps:
wherein, PiRepresenting the degree of preference of a gray-scale image of the cloth of level i, ai、biRespectively representing the width and length of a periodic region on a gray-scale image of a piece of cloth of level i, ai+1、bi+1Respectively representing the width and length of a period region on a gray-scale image of the cloth with the level of i +1, (a)i*bi)－(ai+1*bi+1) The method comprises the steps that the difference between the areas of periodic regions in a piece gray-scale image with the grade i and a piece gray-scale image with the grade i +1 is represented, and the smaller the area difference value is, the greater the preference degree of the piece gray-scale image with the grade i is; the smaller the difference in area, the greater the preference. A. theFiTexture complexity, A, for a piece goods gray-scale image of level iFMRepresenting the maximum texture complexity. A. theFi－AFMAnd expressing the difference value between the texture complexity of the piece gray-scale image with the grade i and the maximum texture complexity, wherein the smaller the difference value is, the greater the preference degree is.
It should be noted that, as the level increases, the details of the cloth texture in the cloth gray-scale image become clearer and the complexity of the texture becomes higher and higher; finally, in a certain level and the subsequent cloth gray scale images, the complexity of the texture of the cloth is almost the same, and the texture cannot be obviously increased; that is to say, the texture information in the cloth gray-scale images is almost the same, and then the gray-scale co-occurrence matrix is generated by selecting the image with the smallest level from the cloth gray-scale images, so that the texture of the current cloth can be accurately represented, and the texture features of the cloth cannot be extracted due to the excessive calculation amount of the gray-scale co-occurrence matrix caused by the excessive number of the grades.
Finally, acquiring a cloth gray-scale image of a grade corresponding to the maximum value of the optimal degree, and recording the image as an optimal cloth image; and calculating the defect rate of the cloth according to the gray level co-occurrence matrix of the optimal cloth image, wherein the cloth with the defect rate larger than a set threshold value is defective.
Specifically, a cloth gray-scale image of a level corresponding to the maximum value of the optimal degree is obtained and recorded as an optimal cloth image. Calculating the defect rate of the current cloth for the optimal cloth image:
acquiring a gray level co-occurrence matrix of each period region in the optimal cloth image, calculating an entropy value of the gray level co-occurrence matrix of each period region, D (ENT) representing the variance of entropy values of all period regions, and E (ENT) representing the mean value of entropy values of all navicular regions.
Setting a threshold XT(threshold value X in the present embodiment)TThe value is 0.3): if X is less than XTIndicating that the current cloth has no defects; if X is more than or equal to XTIndicating that the current piece of cloth is defective. It should be noted that the texture in the same period region is the same, and once the variance is too large, it indicates that the texture in different period regions is different, i.e. it is defective.
It should be noted that: the precedence order of the above embodiments of the present invention is only for description, and does not represent the merits of the embodiments. And specific embodiments thereof have been described above. Other embodiments are within the scope of the following claims. In some cases, the actions or steps recited in the claims may be performed in a different order than in the embodiments and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing may also be possible or may be advantageous.
The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.
Claims (8)
1. A cloth defect detection method based on artificial intelligence is characterized by comprising the following steps:
dividing the gray level image into different grades according to the gray level range of the gray level image of the cloth to obtain the gray level image of the cloth with different grades;
acquiring frequency spectrum information of any piece of cloth gray-scale image, and performing region division on the piece of cloth gray-scale image according to the frequency spectrum information to obtain a plurality of periodic regions;
obtaining the gradient value and coordinates of each pixel point in a period region to form label data of the period region, and classifying the label data to obtain a plurality of classes, wherein each class corresponds to a local region;
acquiring the edge complexity and the internal complexity of each local area, and acquiring the texture complexity of the periodic area according to the edge complexity, the internal complexity and the area of all the local areas;
calculating the sum of the texture complexity of all periodic regions on the cloth gray-scale image to obtain the texture complexity of the cloth gray-scale image, and acquiring the maximum texture complexity in all levels of the cloth gray-scale image;
obtaining the optimal degree of the cloth gray-scale image at the current level according to the difference between the areas of the periodic regions in the cloth gray-scale image at the current level and the cloth gray-scale image at the adjacent level and the difference between the texture complexity of the cloth gray-scale image at the current level and the maximum texture complexity;
acquiring a cloth gray scale image of a grade corresponding to the maximum value of the optimal degree, and recording the image as an optimal cloth image; and calculating the defect rate of the cloth according to the optimal cloth image, wherein the cloth with the defect rate larger than a set threshold value is defective.
2. The cloth defect detection method based on artificial intelligence as claimed in claim 1, wherein said area division of the cloth gray scale image is specifically:
acquiring frequency spectrum information of any piece of cloth gray-scale image to obtain a gray-scale amplitude-frequency image, and establishing an amplitude-frequency coordinate system on the gray-scale amplitude-frequency image by taking the central point of the image as an original point, the transverse direction as an x axis and the longitudinal direction as a y axis;
respectively selecting maximum points on an x axis and a y axis as a transverse periodic point and a longitudinal periodic point; obtaining the width and the length according to the reciprocal of the frequency of the transverse periodic points and the longitudinal periodic points; and dividing the cloth gray scale image into a plurality of periodic areas according to the width and the length.
3. The cloth defect detection method based on artificial intelligence as claimed in claim 1, wherein said dividing different levels according to the gray value range of the cloth gray image is specifically:
and all 256 divisible integers in the gray value range of the cloth gray image are used as division levels, and the cloth gray image is divided into different levels.
4. The cloth defect detection method based on artificial intelligence as claimed in claim 1, wherein the method for obtaining the edge complexity of the local area specifically is:
wherein,
the edge complexity of the local area, N is the number of the edge pixel points of the local area; calculating the hessian matrix of each edge pixel point in the local area,
representing the eigenvector corresponding to the maximum eigenvalue of the Hessian matrix of the nth edge pixel point,
representing the eigenvector corresponding to the maximum eigenvalue of the Hessian matrix of the n-1 th edge pixel point,
to represent
And
similarity of two feature vectors.
5. The cloth defect detection method based on artificial intelligence as claimed in claim 1, wherein the method for obtaining the internal complexity of the local area specifically is:
wherein A isOThe internal complexity of the local area, and M is the number of internal pixel points of the local area;
representing the gradient value of the mth pixel point in the local area,
representing the gradient value of the m-1 th pixel point in the local area,
representing the similarity of the gradient values of the two pixels.
6. The cloth defect detection method based on artificial intelligence as claimed in claim 1, wherein the method for obtaining the texture complexity of the periodic region specifically is:
wherein A is the texture complexity of the periodic region,
indicating the edge complexity of the pth local area,
representing the internal complexity, S, of the p-th local areapThe area of the p-th local region is shown.
7. The cloth defect detection method based on artificial intelligence as claimed in claim 1, wherein the preferred degree obtaining method specifically comprises:
wherein,
representing the degree of preference of a gray-scale image of the cloth of level i, ai、biRespectively representing the width and length of a periodic region in a gray-scale image of a piece of cloth of level ii+1、bi+1Respectively representing the width and length of a period region on a gray-scale image of the cloth with the level of i +1, (a)i*bi)－(ai+1*bi+1) Indicates the difference between the areas of the periodic regions in the gray-scale image of the piece at the level i and the gray-scale image of the piece at the level i +1, AFiRepresenting a gray level of a piece of cloth of level iTexture complexity of the image, AFMRepresenting the maximum texture complexity.
8. The cloth defect detection method based on artificial intelligence as claimed in claim 1, wherein the method for obtaining the defect rate of the cloth is specifically as follows:
acquiring a gray level co-occurrence matrix of each period region in the optimal cloth image, calculating an entropy value of the gray level co-occurrence matrix of each period region, D (ENT) representing the variance of the entropy values of all period regions, and E (ENT) representing the mean value of the entropy values of all period regions.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210401115.6A CN114494259B (en) | 2022-04-18 | 2022-04-18 | Cloth defect detection method based on artificial intelligence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210401115.6A CN114494259B (en) | 2022-04-18 | 2022-04-18 | Cloth defect detection method based on artificial intelligence |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114494259A true CN114494259A (en) | 2022-05-13 |
| CN114494259B CN114494259B (en) | 2022-06-17 |
Family
ID=81489284
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210401115.6A Active CN114494259B (en) | 2022-04-18 | 2022-04-18 | Cloth defect detection method based on artificial intelligence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114494259B (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114663435A (en) * | 2022-05-25 | 2022-06-24 | 启东新朋莱纺织科技有限公司 | Automatic monitoring method for tension abnormity of textile fabric |
| CN114821284A (en) * | 2022-06-30 | 2022-07-29 | 南通捷茜纺织科技有限公司 | Intelligent adjusting method of plaiter for textile production |
| CN114882026A (en) * | 2022-07-08 | 2022-08-09 | 山东科尔自动化仪表股份有限公司 | Sensor shell defect detection method based on artificial intelligence |
| CN115115637A (en) * | 2022-08-30 | 2022-09-27 | 南通市昊逸阁纺织品有限公司 | Cloth defect detection method based on image pyramid thought |
| CN115294137A (en) * | 2022-10-09 | 2022-11-04 | 南通市通州区欢伴纺织品有限公司 | Cloth surface color bleeding defect detection method |
| CN115330766A (en) * | 2022-10-12 | 2022-11-11 | 南通新诚电子有限公司 | Method for identifying formation quality defects of capacitor anode aluminum foil |
| CN116246174A (en) * | 2023-04-26 | 2023-06-09 | 山东金诺种业有限公司 | Sweet potato variety identification method based on image processing |
| CN116559183A (en) * | 2023-07-11 | 2023-08-08 | 钛玛科（北京）工业科技有限公司 | Method and system for improving defect judging efficiency |
| CN116740056A (en) * | 2023-08-10 | 2023-09-12 | 梁山水泊胶带股份有限公司 | Defect detection method for coating layer of whole-core high-pattern conveyer belt |
| CN117152129A (en) * | 2023-10-27 | 2023-12-01 | 智翼博智能科技(苏州)有限公司 | Visual detection method and system for surface defects of battery cover plate |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100158373A1 (en) * | 2008-12-18 | 2010-06-24 | Dalong Li | Methods and apparatus for auto image binarization |
| CN107870172A (en) * | 2017-07-06 | 2018-04-03 | 黎明职业大学 | A kind of Fabric Defects Inspection detection method based on image procossing |
| CN109934802A (en) * | 2019-02-02 | 2019-06-25 | 浙江工业大学 | A kind of Fabric Defects Inspection detection method based on Fourier transformation and morphological image |
-
2022
- 2022-04-18 CN CN202210401115.6A patent/CN114494259B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100158373A1 (en) * | 2008-12-18 | 2010-06-24 | Dalong Li | Methods and apparatus for auto image binarization |
| CN107870172A (en) * | 2017-07-06 | 2018-04-03 | 黎明职业大学 | A kind of Fabric Defects Inspection detection method based on image procossing |
| CN109934802A (en) * | 2019-02-02 | 2019-06-25 | 浙江工业大学 | A kind of Fabric Defects Inspection detection method based on Fourier transformation and morphological image |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114663435A (en) * | 2022-05-25 | 2022-06-24 | 启东新朋莱纺织科技有限公司 | Automatic monitoring method for tension abnormity of textile fabric |
| CN114821284A (en) * | 2022-06-30 | 2022-07-29 | 南通捷茜纺织科技有限公司 | Intelligent adjusting method of plaiter for textile production |
| CN114882026A (en) * | 2022-07-08 | 2022-08-09 | 山东科尔自动化仪表股份有限公司 | Sensor shell defect detection method based on artificial intelligence |
| CN114882026B (en) * | 2022-07-08 | 2022-09-02 | 山东科尔自动化仪表股份有限公司 | Sensor shell defect detection method based on artificial intelligence |
| CN115115637B (en) * | 2022-08-30 | 2022-12-06 | 南通市昊逸阁纺织品有限公司 | Cloth defect detection method based on image pyramid thought |
| CN115115637A (en) * | 2022-08-30 | 2022-09-27 | 南通市昊逸阁纺织品有限公司 | Cloth defect detection method based on image pyramid thought |
| CN115294137A (en) * | 2022-10-09 | 2022-11-04 | 南通市通州区欢伴纺织品有限公司 | Cloth surface color bleeding defect detection method |
| CN115330766B (en) * | 2022-10-12 | 2023-04-11 | 南通新诚电子有限公司 | Method for identifying formation quality defects of capacitor anode aluminum foil |
| CN115330766A (en) * | 2022-10-12 | 2022-11-11 | 南通新诚电子有限公司 | Method for identifying formation quality defects of capacitor anode aluminum foil |
| CN116246174A (en) * | 2023-04-26 | 2023-06-09 | 山东金诺种业有限公司 | Sweet potato variety identification method based on image processing |
| CN116246174B (en) * | 2023-04-26 | 2023-08-08 | 山东金诺种业有限公司 | Sweet potato variety identification method based on image processing |
| CN116559183A (en) * | 2023-07-11 | 2023-08-08 | 钛玛科（北京）工业科技有限公司 | Method and system for improving defect judging efficiency |
| CN116559183B (en) * | 2023-07-11 | 2023-11-03 | 钛玛科（北京）工业科技有限公司 | Method and system for improving defect judging efficiency |
| CN116740056A (en) * | 2023-08-10 | 2023-09-12 | 梁山水泊胶带股份有限公司 | Defect detection method for coating layer of whole-core high-pattern conveyer belt |
| CN116740056B (en) * | 2023-08-10 | 2023-11-07 | 梁山水泊胶带股份有限公司 | Defect detection method for coating layer of whole-core high-pattern conveyer belt |
| CN117152129A (en) * | 2023-10-27 | 2023-12-01 | 智翼博智能科技(苏州)有限公司 | Visual detection method and system for surface defects of battery cover plate |
| CN117152129B (en) * | 2023-10-27 | 2024-02-13 | 智翼博智能科技(苏州)有限公司 | Visual detection method and system for surface defects of battery cover plate |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114494259B (en) | 2022-06-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114494259B (en) | Cloth defect detection method based on artificial intelligence | |
| CN113450307B (en) | Product edge defect detection method | |
| CN109829914B (en) | Method and device for detecting product defects | |
| CN108229526B (en) | Network training method, network training device, image processing method, image processing device, storage medium and electronic equipment | |
| JP3877916B2 (en) | Anomaly detection method and system for digital image, and storage medium therefor | |
| JP4590471B2 (en) | Method and system for estimating background color | |
| CN115345885A (en) | Method for detecting appearance quality of metal fitness equipment | |
| CN111626993A (en) | Image automatic detection counting method and system based on embedded FEFnet network | |
| CN113554629A (en) | Strip steel red rust defect detection method based on artificial intelligence | |
| CN115082451B (en) | Stainless steel soup ladle defect detection method based on image processing | |
| CN109035196B (en) | Saliency-based image local blur detection method | |
| CN113888536B (en) | Printed matter double image detection method and system based on computer vision | |
| CN111652213A (en) | Ship water gauge reading identification method based on deep learning | |
| CN114998314B (en) | PCB defect detection method based on computer vision | |
| CN111598869B (en) | Method, equipment and storage medium for detecting Mura of display screen | |
| CN110648330B (en) | Defect detection method for camera glass | |
| CN114782329A (en) | Bearing defect damage degree evaluation method and system based on image processing | |
| CN111784675A (en) | Method and device for processing article texture information, storage medium and electronic equipment | |
| CN111972700A (en) | Cigarette appearance detection method and device, equipment, system and medium thereof | |
| CN115330795A (en) | Cloth burr defect detection method | |
| CN114757913A (en) | Display screen defect detection method | |
| CN117094975A (en) | Method and device for detecting surface defects of steel and electronic equipment | |
| CN113609984A (en) | Pointer instrument reading identification method and device and electronic equipment | |
| CN114519694B (en) | Seven-segment nixie tube liquid crystal display screen identification method and system based on deep learning | |
| CN117689550A (en) | Low-light image enhancement method and device based on progressive generation countermeasure network |
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 |