Defect detection in coloured texture surfaces using Gabor filters

The present paper presents a Gabor filtering approach for the automatic inspection of defects in coloured texture surfaces. It can simultaneously measure both chromatic and textural anomalies in an image. Two chromatic features derived from the CIE-L*a*b* colour space are used to form a complex number for colour pixel representation. The proposed method is based on the energy response from the convolution of a Gabor filter with the colour image characterized by two chromatic features in the form of a complex number. The Gabor filtering process converts the difficult defect detection in a coloured texture image into simple threshold segmentation in the filtered image. Experimental results from a number of coloured texture surfaces such as textile fabric, wood and the have shown the effectiveness of the proposed method.