Realization of Orthogonal Wavelets Adapted to Fabric Texture for Defect Detection

摘要

The wavelet adapted to the fabric texture can be developed from the orthogonal and normal series which are selected randomly by means of Monte Carlo method and optimized by adding certain constraint conditions. Then the fabric image can be decomposed into the subimages by the adaptive wavelet transform and the horizontal and vertical texture information will be perfectly contained in the subimages. Therefore this method can be effectively used for the automatic inspection of the fabric defects.