Fabric defect detection with Gabor filter based on scale transformation

Focusing on automatic image inspection of fabric defects in textile industry,a method based on scale transformation Gabor filter was presented,in which the expansion scale varies with the center frequency in Gabor filter function.The filter groups of three scales,towards four directions,take the image of normal fabric and that of the fabric to be identified,extract their texture features,and thus obtain the deviation images that are not conformed with those of the normal fabric.Integration of the deviation images make the defects of the fabric prominent.And finally,fabric defects are detected by threshold processing.The experimental results of four typical defects including missing warp,missing weft,error interlacing,and oil stain prove that this method is effective.