A Texture Defect Feature Extraction Algorithm
WU Huanxin
YUE Xiaofeng
QIN Weiyang
ZHANG Pengfei
Abstract：In order to extract the fine surface defects of the object with texture more accurately,this paper presents a feature extraction algorithm for texture surface defects based on the second Curvelet transform,gray level co-occurrence matrix and second order oscillation particle swarm optimization. Firstly,the texture defect image is decomposed on different scales and directions,and the high frequency components containing the texture features are found by using the gray level co-occurrence matrix. The high-fre?quency components are extracted and the high-frequency components of the defect details are extracted. Finally,the similarity func?tion is used to find the similarity between the eigenvector of the test sample and the eigenvector of the training sample. According to the similarity value of the test sample,the similarity degree of the test sample is obtained by using the kernel principal component analysis algorithm. Size determines the type of defect for the test sample. The results of this algorithm are more complete and the rec?ognition rate of the defect is higher than that of the non-optimized kernel principal component analysis feature extraction algorithm and the kernel principal component analysis feature extraction algorithm based on clustering.
Keywords：second Curvelet transformgray covariance matrixkernel principal component analysissecond- order oscilla?tion particle swarm optimizationdefect feature
Publication Date：2019-01-01
Online Publishing Date：2025-08-15(First online date of this platform, not the publication date of the document)
Pages：6( 1055-1059,1077 )