The inspection of fabric defects based on wavelet transform

In the textile production, there may appear many fabric defects. To fabric defects, there are a lot of image-based inspection techniques: Fourier transform, Sobel algorithm of edge inspection, fast Fourier transform (FFT) et. However, Wavelet transform is a kind of multiresolution algorithm, and its multiresolution character corresponds to time frequency multiresolution of human vision. the result of the experiment proves that wavelet transform gives better results than the other traditional methods. So in this article, we use wavelet transform and BP neural network together to inspect and classes the fabric defects.