Inspection of fabric defects based on wavelet analysis and BP neural network

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 research indicates 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 classify the fabric defects. A plain white fabric is adopted as the sample, and the distinguishing defects are oil stains, warp-lacking, and weft-lacking. An area camera with 256times256 resolution is used in the scheme, a grabbed image is transmitted to a computer for wavelet transform, and then the corresponding image data are then used in BP neural network as input. The result shows that the fabric defectspsila classification rate can be up to 95% with above method.