Fault segmentation in fabric images using Gabor wavelet transform

Gabor wavelets have been successfully applied for a variety of machine vision applications such as Texture segmentation, Edge detection, Boundary detection etc. As the Fourier transform is not suitable for detecting local defects, and the Wavelet transforms posses only limited number of orientations, Gabor wavelet transform is chosen and applied to detect the defects in fabrics. Gabor filters scheme that imitates the early human vision process is applied to the sample under inspection. Defects can be automatically segmented from the regular texture by applying the proposed method. Proper thresholding ensures segmentation of the defect from the texture background. The results obtained using this method confirms its efficiency. This can also be applied to detect defects on surfaces and materials that have regular periodic texture.