A Fabric Defect Detection Method Based on Sparse Representation Coefficient Optimization
A technology of sparse representation coefficient and sparse coefficient, which is applied in image data processing, instrumentation, calculation, etc., can solve the problem that the dictionary library is difficult to accurately represent normal texture, the original fabric reconstruction error is large, and the missed detection, etc., to achieve fast calculation speed, Strong adaptability and robustness, and the effect of expanding the range of use
Active Publication Date: 2017-08-04
ZHONGYUAN ENGINEERING COLLEGE
View PDF4 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
This method can reconstruct texture images of different fabrics, and has high self-adaptive characteristics; however, when this technology is applied to pattern fabric dictionary detection, the trained dictionary library is difficult to accurately represent normal textures, and the original fabric reconstruction error is relatively large. Large, resulting in false detection, missed detection, etc.
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment
[0087] In the embodiment, images of common defects in the fabric image library are used for experiments, including images of leaking yarn, damage, loose weft, jumping flowers, knots, etc. The size of the image is 256×256, select some images such as figure 2 (a)-(f) in. In the embodiment, the value of k is 4, the value of λ in the formula (1) is 0.05, and the value of γ in the formula (3) is 1.5. Using the sparse representation coefficient matrix α, α before and after optimization * and the dictionary library D to reconstruct the fabric image to be tested, the results are as follows image 3 (a)-(f) and Figure 4 (a)-(f) in. Depend on image 3 (a)-(f) and Figure 4 (a)-(f) are compared one by one, and the image reconstructed by using the sparse representation coefficient matrix α is image 3 In (a)-(f) there are relatively obvious defect areas, and the optimized sparse representation coefficient matrix α is used * owned Figure 4 The reconstructed images in (a)-(f) ar...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention discloses a fabric defect detection method based on sparse representation coefficient optimization, including self-adaptive dictionary library learning, sparse coefficient matrix optimization, image reconstruction, generation and segmentation of visually salient maps, image segmentation, and automatic Adapt to the dictionary library learning and obtain the dictionary library; use the L2 norm minimization method to obtain the sparse representation coefficient matrix, and optimize the abnormal coefficient elements in the obtained matrix; use the obtained dictionary library and the optimized sparse representation coefficient matrix to reconstruct the fabric image, And make a residual difference with the image to be tested to obtain a residual saliency map; use the maximum entropy threshold segmentation method to segment the saliency map to obtain the fabric defect detection result. The present invention comprehensively considers the randomness of fabric texture features and the diversity of defect types, uses the fabric image to be tested as a dictionary learning sample and a defect area detection reference, and has high detection accuracy; and does not need to extract any defect information, self-adaptive Strong ability; fast calculation speed, suitable for online detection.
Description
technical field [0001] The invention belongs to the technical field of textile image processing, and in particular relates to a method for detecting and locating defects on fabric defect images by using an image sparse representation method and a saliency analysis method. Background technique [0002] Fabric defect detection is a key link in textile quality control and management. With the rapid development of integrated circuits and image processing technology, machine vision has been more and more widely used in the field of industrial surface inspection. Replacing manual operations with computer vision can not only increase the detection speed and reduce labor costs, but also through cloth The defect automatic detection system can provide credible reference standards for both parties for the assessment of cloth quality grades, which is beneficial to international trade. Fabric defect detection and discrimination algorithm is the core link of this kind of system, which di...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & Authority Patents(China)
IPC IPC(8): G06T7/00
Inventor 刘洲峰李春雷董燕闫磊余淼
Owner ZHONGYUAN ENGINEERING COLLEGE