Fabric Defect Identification Using Back Propagation Neural Networks

Abstract

Fabric defect identification plays a very important role for the automatic detection in fabrics. Fabric defect identification mainly includes three parts: The first, preprocessing with Frequency domain Butterworth Low pass Filter and Histogram Equalization. The second, extraction of texture features from fabric using Gray Level Co-occurrence Matrix (GLCM).The Co-occurrence matrix characterizes the distribution of co-occurring pixel values in an image to be at a given offset, and then the statistical features are extracted from this matrix. The Third, the extracted GLCM features are used for the classification of the texture using Back Propagation Neural Network with different learning rules for their effectiveness comparison.