Study of Fabric Defects Detection through Gabor Filter Based on Scale Transformation

摘要

Aiming at fabric defects detection in textile industry, a Gabor filtering strategy based on scale transformation was presented, in which the extension scale of Gabor filter function was shown. Firstly the Gabor filter banks are convolved with normal fabric image to extract normal texture features. Secondly the convolution of Gabor filter banks and defected image is carried out to obtain the deviation images, and then they are fused. Lastly the fabric defects are detected by threshold processing. The results show that the method is effective for fabric defects detecting.