Nep Detection in Carding Machine Using Optimized Filters

In this paper, a new approach for nep detection in carding machine using optimal Gabor filters is proposed. The optimal filters based on the discrete Gabor filters have more quick speed for feature extraction from cotton web samples. 7 x 7 masks are used for computational saving. The nep detection processing is divided into following steps: pre-processing, feature extraction by optimal Gabor filters and feature enhancement, image fusion, image thresholding and segmentation. The nep detection accuracy of experiments is 95.2%. The excellent results confirm the usefulness of the approach for industrial inspection.