CN108765402B - Non-woven fabric defect detection and classification method - Google Patents
Non-woven fabric defect detection and classification method Download PDFInfo
- Publication number
- CN108765402B CN108765402B CN201810535590.6A CN201810535590A CN108765402B CN 108765402 B CN108765402 B CN 108765402B CN 201810535590 A CN201810535590 A CN 201810535590A CN 108765402 B CN108765402 B CN 108765402B
- Authority
- CN
- China
- Prior art keywords
- defect
- image
- woven fabric
- feature vector
- group
- 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
- 230000007547 defect Effects 0.000 title claims abstract description 83
- 239000004745 nonwoven fabric Substances 0.000 title claims abstract description 46
- 238000000034 method Methods 0.000 title claims abstract description 39
- 238000001514 detection method Methods 0.000 title claims abstract description 24
- 239000013598 vector Substances 0.000 claims abstract description 32
- 238000001914 filtration Methods 0.000 claims abstract description 25
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 8
- 238000012549 training Methods 0.000 claims abstract description 7
- 239000002131 composite material Substances 0.000 claims abstract description 6
- 230000011218 segmentation Effects 0.000 claims abstract description 5
- 230000004927 fusion Effects 0.000 claims description 9
- 238000011156 evaluation Methods 0.000 claims description 7
- 238000005457 optimization Methods 0.000 claims description 6
- 230000008030 elimination Effects 0.000 claims description 3
- 238000003379 elimination reaction Methods 0.000 claims description 3
- 238000005286 illumination Methods 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 claims description 2
- 238000013461 design Methods 0.000 abstract description 3
- 239000004744 fabric Substances 0.000 abstract 1
- 238000013528 artificial neural network Methods 0.000 description 3
- 230000002950 deficient Effects 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000012706 support-vector machine Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 230000006978 adaptation Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000003708 edge detection Methods 0.000 description 1
- 238000007499 fusion processing Methods 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 230000004807 localization Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000000275 quality assurance Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000012795 verification 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
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G06T5/70—
-
- G06T5/90—
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- 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
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20004—Adaptive image processing
- G06T2207/20012—Locally adaptive
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20212—Image combination
- G06T2207/20221—Image fusion; Image merging
-
- 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
Abstract
The invention discloses a non-woven fabric defect detection and classification method, which solves the automatic detection and classification problems of four defects of non-woven fabric holes, oil stains, foreign matters and scratches. Firstly, detecting a non-woven fabric defect image, filtering the non-woven fabric defect image by using an optimized Gabor filter group, fusing filtering results, binarizing the non-woven fabric defect image by using a self-adaptive threshold value segmentation method, and eliminating noise interference by using a pseudo defect eliminating algorithm so as to accurately position the position of a defect in the image; secondly, segmenting an interested region in the image according to the position of the defect, and extracting a composite feature vector consisting of shape features, first-order moment features and second-order moment features based on the interested region; training an SVM classifier by using the composite feature vector group and a one-to-one design strategy; and finally, accurately classifying the defect characteristics of the non-woven fabric by using the trained classifier group. The method has the advantages of accurate defect positioning and high classification accuracy, and is used for detecting and classifying the cloth defects of non-woven fabric manufacturers.
Description
Technical Field
The invention belongs to the field of image identification, and particularly relates to a non-woven fabric defect detection and classification method which can be used for identifying defect images acquired in a non-woven fabric quality detection link.
Background
Non-woven fabric defect detection mainly realizes through the manual work at present as an effectual quality assurance means, and work load is big and detection efficiency is not high. Therefore, the automatic machine vision detection is a reasonable choice, and can ensure higher detection speed and detection rate. Aiming at the problem of detection and positioning of defects of non-woven fabrics, a method based on Gabor filtering is mainly used at present. Huazhong science and technology university Liuhaiping et al adopt the multi-direction multiscale Gabor filter to realize the detection location of non-woven fabrics defect, and the rate of accuracy that detects is higher, but the real-time is relatively poor. Zhai Bohai et al, China university of science and technology, adopts a one-way and one-scale Gabor filter to realize the detection and positioning of non-woven fabric defects, the algorithm has good real-time performance, but the defect information loss is large, and the detection accuracy is poor. Aiming at the classification problem of the defects of the non-woven fabric, an artificial neural network-based method is mainly used at present. The Zhai and Bohai, China university of science and technology, and the like, adopt a three-layer BP neural network to automatically classify the defects of the non-woven fabric, the accuracy rate of classification can reach 87.05%, but the neural network algorithm has more requirements on the number of training samples and is difficult to realize.
Disclosure of Invention
An object of an aspect of the present disclosure is to provide a method for detecting and classifying defects of a non-woven fabric, so as to quickly locate a defect region in a defect image of the non-woven fabric and accurately classify types of the defects, thereby realizing automation of quality detection of the non-woven fabric. The method specifically comprises the following steps: acquiring a non-woven fabric image; carrying out brightness compensation and filtering denoising pretreatment on the non-woven fabric image so as to eliminate the influence of uneven illumination on the image and filter noise generated in the image acquisition process; filtering the preprocessed image by adopting an optimized Gabor filter group, and fusing the filtered image to enable defect information to be concentrated in a fusion result; performing binarization on the fusion result graph by adopting a local mean value self-adaptive threshold segmentation method to separate a defect region from a normal region; removing isolated bright noise points in the binarization result image by adopting a pseudo-defect elimination algorithm, so that only a correct defect area is reserved in the binarization image, and the detection and positioning of the defects are accurately realized; dividing an interested region according to the position of the defect in the image, and extracting defect characteristics based on the interested region; training an SVM classifier according to the defect feature vector group and carrying out parameter optimization; and classifying and identifying the defects by using the trained classifier.
In the above method for detecting and classifying defects of a non-woven fabric, the method for optimizing the Gabor filter bank comprises the following steps: designing 40 Gabor filters in the directions of 5 scales and 8 in total according to a two-dimensional Gabor function; dividing 8 directions into 4 groups of orthogonal directions, calculating the sum of one-dimensional entropies of filtering result images in each group of orthogonal directions according to a calculation method of the one-dimensional entropies of the images, and selecting the orthogonal direction group corresponding to the minimum value as the optimal direction of the Gabor filter group; calculating loss evaluation function values of the filtering result images of all scales in the optimal direction, calculating the sum of the loss evaluation function values in each group of orthogonal directions, and selecting the scale corresponding to the maximum value as the optimal scale of the Gabor filter group; the method comprises the steps of setting the direction parameters of the Gabor filter to be the optimal direction, setting the scale parameters to be the optimal scale, only taking a part of the Gabor filter to complete the design of an optimized Gabor filter group, and filtering a non-woven fabric image by using the optimized Gabor filter group.
In the method for detecting and classifying the defects of the non-woven fabric, the filtering result graphs in two orthogonal directions are weighted and fused, so that the information of the defects is concentrated in the fused result graph.
In the above method for detecting and classifying defects of non-woven fabric, the defect features include shape feature vectors, first moment feature vectors and second moment feature vectors, and the shape feature vectors, the first moment feature vectors and the second moment feature vectors are combined into a composite feature vector as a feature representing a defect type.
In the above method for detecting and classifying defects of non-woven fabric, the shape feature vector includes parameters: area, perimeter, orientation angle, circularity, flatness, and duty cycle.
In the above method for detecting and classifying defects of non-woven fabric, the first moment eigenvector includes parameters: gray level mean, variance, slope, kurtosis, and one-dimensional entropy.
In the above method for detecting and classifying defects of non-woven fabric, the second-order moment eigenvector includes parameters: energy, two-dimensional entropy, contrast, inverse difference moment, and autocorrelation over the optimal set of orthogonal directions.
The beneficial effects of the invention are as follows:
1. according to the method, the texture characteristics of the non-woven fabric and the evaluation standard of the related image are combined, 2 Gabor filters in the 1-scale 2 direction are selected from 40 Gabor filters in the 5-scale 8 direction to form an optimized Gabor filter group, and the optimized Gabor filter group greatly reduces the running time of the algorithm on the premise of ensuring the accuracy of defect detection, so that the algorithm has higher practicability.
2. The invention adopts the support vector machine to classify, optimizes the parameters of the classifier and can obtain higher classification accuracy under the condition of less training samples.
Drawings
FIG. 1 is a flow chart of a method for detecting and classifying defects in a nonwoven fabric according to an embodiment of the present invention.
FIG. 2 is a sample of 4 types of common nonwoven defects collected by the present invention; in the figure, (a) is a hole sample, (b) is a foreign matter sample, (c) is an oil stain sample, and (d) is a scratch sample.
FIG. 3 is a diagram of simulation results of defect localization for a 4-class common non-woven fabric defect sample test set according to the present invention; in the figure, (a) is a result of positioning a hole breaking defect, (b) is a result of positioning a foreign matter defect, (c) is a result of positioning an oil stain defect, and (d) is a result of positioning a scratch defect.
FIG. 4 is a flow diagram of classifier parameter optimization according to one embodiment of the present invention.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings and specific embodiments.
In order to quickly locate the defect area in the non-woven fabric defect image and accurately classify the defect type, the invention provides a non-woven fabric defect automatic detection and classification method, and the automation of non-woven fabric quality detection is realized.
Step 1, acquiring a gray level image of the non-woven fabric from a non-woven fabric detection production line, wherein the image can be acquired by a common industrial camera. The defect pictures are selected, and 4 defects of holes, oil stains, foreign matters and scratches are collected, as shown in fig. 2, each defect can comprise 50 pictures, all the pictures can be 8-bit gray-scale pictures, and the size can be 1280 x 960. Each class may use 80% of the pictures randomly selected as training samples and all pictures may be used as test samples.
And 2, performing brightness compensation and filtering denoising on the non-woven fabric gray level image to eliminate the influence of uneven illumination on the image and effectively filter noise generated in the image acquisition process. The brightness compensation can be carried out on the non-woven fabric image by adopting a block mean value method, and the filtering and de-noising can be carried out on the non-woven fabric image by adopting a median filter with the size of 3 multiplied by 3.
And 3, filtering the preprocessed non-woven fabric image by adopting an optimized Gabor filter bank, wherein the optimization process of the Gabor filter bank can refer to the steps 3a to 3 c.
Step 3a, generating 40 Gabor filters in the direction of 5 scales and 8 directions according to a formula (1);
the Gabor filtering can be divided into a real part filtering and an imaginary part filtering, and can be referred to formulas (2) and (3):
the real part of the Gabor filter has a Gaussian smoothing effect, the imaginary part of the Gabor filter has an edge detection effect, and the defect detection of the non-woven fabric image mainly uses the effect of partial filtering of the non-woven fabric image, so that the image is filtered only by the real part of the Gabor filter.
Step 3b, divide 8 directions into 4 sets of orthogonal directions, i.e. 0 ° and 90 °, 22.5 ° and 112.5 °, 45 ° and 135 °, 77.5 ° and 157.5 °, calculate the one-dimensional entropy of the image for the filtered result map in each direction:
and adding the values of the one-dimensional entropy of the filtering result images in each group of orthogonal directions, and selecting the group of orthogonal directions with the minimum sum as the optimal direction group of the Gabor filter group.
Step 3c, setting the direction of the Gabor filter group as an optimal direction group, taking Gabor filters corresponding to 5 scales in each direction to filter the image, and calculating a loss evaluation function value of each filtering result image:
and adding the loss evaluation function values of the filtering result images in each group of orthogonal directions, and selecting the maximum sum scale as the optimal scale selection of the Gabor filter group. The optimized Gabor filter bank obtained finally is GRaAnd GRbThe corresponding filtering result is IaAnd Ib。
Step 4, mixing IaAnd IbAnd performing image fusion to centralize the defect information into a fusion result, wherein the image fusion can be realized by adopting a weighted fusion algorithm, and the fusion process can be expressed as follows:
Rm=αIa+βIb,α+β=1 (6);
in the formula (6), RmIs a fusion result graph, due to IaAnd IbIs orthogonal, so α β is 0.5.
Local mean adaptation can be used to determine the threshold for each pixel segmentation,
T(i,j)=Wk*Rm(i,j)-C (7);
the binarization process can be expressed as:
thereby obtaining a binarized image R showing the defective positionst。
Step 6, the binary image R is processedtAnd performing false defect eliminating operation to prevent some isolated bright noise points in the binary image from influencing the positioning of the defects.
A connected domain based approach may be employed to reject false defects. In a first step, R istThe bright pixel points in (1) are labeled as label1Dark pixel point markIs marked as label2(ii) a Second, for the image RtTraversing is carried out, and when the mark of the pixel point (x, y) is label1Then, the pixel point is taken as a connected domain CiI ═ 1, 2.. and n; thirdly, scanning the 8 neighborhood pixels of (x, y), and marking the pixels as label1Sequentially adding C to the pixel pointsiAnd will then CiThe labels of all the pixel points in the system are changed into label3(ii) a The fourth step, with CiThe next pixel point in the process is taken as a seed pixel point, the third step is executed in a circulating way, and when C isiWhen no new pixel point is added, the connected region C is completediIs searched, the area size S is recordedi1,2, ·, n; the fifth step, the second step is executed circularly, when the whole image R is finishedtAfter traversal, R can be obtainedtAll connected domains C in1,C2,...,CnThe corresponding area is S1,S2,...,Sn(ii) a Sixthly, setting the rejection threshold value as TaTo make the area smaller than TaThe pixel points of the connected domain are changed into dark pixels, and the elimination of the pseudo defects is completed. The binarized image after the pseudo-defect removal operation is completed is shown in fig. 3, in which a white area is a defective area and a black area is a normal area.
And 7, after the accurate position of the defect in the image is positioned, segmenting the region of interest, extracting defect features aiming at the region of interest so as to improve the speed and the accuracy of classification, for example, extracting a shape feature vector, a first-order moment feature vector and a second-order moment feature vector aiming at the region of interest, and combining the shape feature vector, the first-order moment feature vector and the second-order moment feature vector into a composite feature vector serving as the feature representing the defect type. The shape feature vector includes 6 parameters of area, perimeter, orientation angle, circularity, flatness, and duty cycle. The first moment feature vector includes 5 parameters of gray level mean, variance, inclination, kurtosis, and one-dimensional entropy. The second-order moment eigenvector includes 5 parameters of energy, two-dimensional entropy, contrast, inverse difference moment and autocorrelation in the optimal orthogonal direction set.
Step 8, training SVM classifiers by using the composite feature vector group and performing parameter optimization, wherein 6 classifiers can be trained by adopting a one-to-one design strategy, and the number of the classifiers can be obtained by the following formula (9):
n=k(k-1)/2 (9);
in equation (9), n is the number of classifiers, and k is the number of defect classes. The kernel function of the support vector machine selects a radial basis kernel function, the values of the parameters C and g in the radial basis kernel function have larger influence on the accuracy rate of classification, and the values of the parameters C and g are optimized by adopting a grid search method and a K-fold cross verification method.
And 9, classifying the defect feature vectors by using the trained SVM classifier group, sequentially inputting the defect feature vectors into the trained SVM classifiers, giving a classification result by each classifier, and taking the class with the largest occurrence frequency in the classification result as a final classification result.
Claims (6)
1. A non-woven fabric defect detection and classification method is characterized by comprising the following steps:
acquiring a non-woven fabric image;
carrying out brightness compensation and filtering denoising pretreatment on the non-woven fabric image so as to eliminate the influence of uneven illumination on the image and filter noise generated in the image acquisition process;
performing filtering operation on the preprocessed image by adopting an optimized Gabor filter group, and fusing the filtered image to enable defect information to be concentrated in a fusion result, wherein the Gabor filter group optimization method comprises the following steps: designing 40 Gabor filters in the directions of 5 scales and 8 in total according to a two-dimensional Gabor function; dividing 8 directions into 4 groups of orthogonal directions, calculating the sum of one-dimensional entropies of filtering result images in each group of orthogonal directions according to a calculation method of the one-dimensional entropies of the images, and selecting the orthogonal direction group corresponding to the minimum value as the optimal direction of the Gabor filter group; calculating loss evaluation function values of the filtering result images of all scales in the optimal direction, calculating the sum of the loss evaluation function values in each group of orthogonal directions, and selecting the scale corresponding to the maximum value as the optimal scale of the Gabor filter group; setting the direction parameter of the Gabor filter as the optimal direction, setting the scale parameter as the optimal scale, and taking only the real part of the filter;
performing binarization on the fusion result graph by using a local mean value self-adaptive threshold segmentation method to separate a defect region from a normal region;
removing isolated bright noise points in the binarization result image by adopting a pseudo-defect elimination algorithm, so that only a correct defect area is reserved in the binarization image, and the detection and positioning of the defects are accurately realized;
dividing an interested region according to the position of the defect in the image, and extracting defect characteristics based on the interested region;
training an SVM classifier according to the defect feature vector group and carrying out parameter optimization;
and classifying and identifying the defects by using the trained classifier.
2. The method according to claim 1, wherein the filtering result graphs in two orthogonal directions are weighted and fused to gather the defect information into the fused result graph.
3. The method as claimed in claim 1, wherein the defect features include a shape feature vector, a first moment feature vector and a second moment feature vector, and the shape feature vector, the first moment feature vector and the second moment feature vector are combined into a composite feature vector as the feature representing the defect type.
4. The method as claimed in claim 3, wherein the shape feature vector comprises parameters: area, perimeter, orientation angle, circularity, flatness, and duty cycle.
5. The method as claimed in claim 3, wherein the first moment eigenvector comprises parameters: gray level mean, variance, slope, kurtosis, and one-dimensional entropy.
6. The method as claimed in claim 3, wherein the second-order moment eigenvector comprises parameters of: energy, two-dimensional entropy, contrast, inverse difference moment, and autocorrelation over the optimal set of orthogonal directions.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810535590.6A CN108765402B (en) | 2018-05-30 | 2018-05-30 | Non-woven fabric defect detection and classification method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810535590.6A CN108765402B (en) | 2018-05-30 | 2018-05-30 | Non-woven fabric defect detection and classification method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108765402A CN108765402A (en) | 2018-11-06 |
CN108765402B true CN108765402B (en) | 2021-09-14 |
Family
ID=64003857
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810535590.6A Active CN108765402B (en) | 2018-05-30 | 2018-05-30 | Non-woven fabric defect detection and classification method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108765402B (en) |
Families Citing this family (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109829906A (en) * | 2019-01-31 | 2019-05-31 | 桂林电子科技大学 | It is a kind of based on the workpiece, defect of the field of direction and textural characteristics detection and classification method |
CN109934811B (en) * | 2019-03-08 | 2022-12-30 | 中国科学院光电技术研究所 | Optical element surface defect detection method based on deep learning |
CN109978844A (en) * | 2019-03-15 | 2019-07-05 | 维库(厦门)信息技术有限公司 | One kind being based on the modular intelligent analysis method of deep learning and system |
CN111127383A (en) * | 2019-03-15 | 2020-05-08 | 杭州电子科技大学 | Digital printing online defect detection system and implementation method thereof |
CN110111711A (en) * | 2019-04-30 | 2019-08-09 | 京东方科技集团股份有限公司 | The detection method and device of screen, computer readable storage medium |
CN110060253B (en) * | 2019-05-06 | 2021-04-20 | 西安交通大学 | Composite sleeper pore defect identification method based on Gabor multi-feature extraction and optimization |
CN112364927A (en) * | 2020-11-17 | 2021-02-12 | 哈尔滨市科佳通用机电股份有限公司 | Foreign matter detection method based on filter bank |
CN114998189A (en) * | 2022-04-15 | 2022-09-02 | 电子科技大学 | Color display point defect detection method |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101286195A (en) * | 2008-06-03 | 2008-10-15 | 西南交通大学 | High precision palm print recognition method based on two-dimensional quadrature Log-Gabor filtering |
CN101866427A (en) * | 2010-07-06 | 2010-10-20 | 西安电子科技大学 | Method for detecting and classifying fabric defects |
CN108090494A (en) * | 2017-12-15 | 2018-05-29 | 东华大学 | Based on Gabor filter and support vector machines textile flaw recognition methods |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP6537376B2 (en) * | 2015-07-01 | 2019-07-03 | キヤノン株式会社 | Image processing apparatus and image processing method |
-
2018
- 2018-05-30 CN CN201810535590.6A patent/CN108765402B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101286195A (en) * | 2008-06-03 | 2008-10-15 | 西南交通大学 | High precision palm print recognition method based on two-dimensional quadrature Log-Gabor filtering |
CN101866427A (en) * | 2010-07-06 | 2010-10-20 | 西安电子科技大学 | Method for detecting and classifying fabric defects |
CN108090494A (en) * | 2017-12-15 | 2018-05-29 | 东华大学 | Based on Gabor filter and support vector machines textile flaw recognition methods |
Also Published As
Publication number | Publication date |
---|---|
CN108765402A (en) | 2018-11-06 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108765402B (en) | Non-woven fabric defect detection and classification method | |
CN109934802B (en) | Cloth defect detection method based on Fourier transform and image morphology | |
CN105784713B (en) | Sealing ring detection method of surface flaw based on machine vision | |
WO2018107939A1 (en) | Edge completeness-based optimal identification method for image segmentation | |
CN106780486B (en) | Steel plate surface defect image extraction method | |
CN107437243B (en) | Tire impurity detection method and device based on X-ray image | |
CN109685766B (en) | Cloth flaw detection method based on region fusion characteristics | |
CN108648168A (en) | IC wafer surface defects detection methods | |
CN108090494B (en) | Textile defect identification method based on Gabor filter and support vector machine | |
CN109816644A (en) | A kind of bearing defect automatic checkout system based on multi-angle light source image | |
CN109685760B (en) | MATLAB-based SLM powder bed powder laying image convex hull depression defect detection method | |
CN108364291A (en) | Grey cloth rapid detection method based on computer vision technique | |
CN107220649A (en) | A kind of plain color cloth defects detection and sorting technique | |
CN108830832A (en) | A kind of plastic barrel surface defects detection algorithm based on machine vision | |
US11538261B2 (en) | Systems and methods for automated cell segmentation and labeling in immunofluorescence microscopy | |
CN107610104A (en) | Crack detecting method at a kind of elevator compensation chain R based on machine vision | |
CN109472788B (en) | Method for detecting flaw on surface of airplane rivet | |
CN106780464A (en) | A kind of fabric defect detection method based on improvement Threshold segmentation | |
CN112614062A (en) | Bacterial colony counting method and device and computer storage medium | |
CN112734761B (en) | Industrial product image boundary contour extraction method | |
Daniel et al. | Automatic road distress detection and analysis | |
CN113221881B (en) | Multi-level smart phone screen defect detection method | |
CN110310291A (en) | A kind of rice blast hierarchy system and its method | |
Ji et al. | Apple grading method based on features of color and defect | |
CN110415222A (en) | A kind of spinning cake side face defects recognition methods based on textural characteristics |
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 |