Abstract:

By analyzing the characteristics of directional fabric textures, a new method for

defect detection is presented. Firstly, the main lines of texture fabric texture are detected

by the normal fabric texture Hough transform. Then, fabric texture images are decomposed by

the directional wavelet. On this basis, the subwindows of characteristics are extracted from

the detailed subimages. Finally, the fabric defects are identified by the BP neural network.

The experimental results show that this method is effective.

Key words: directional texture;Hough transform;wavelet decomposition;defect detection