Intelligent detection of defects of yarn-dyed fabrics by energy-based local binary patterns

For the purpose of realizing fast and effective detection of defects of yarn-dyed fabric via computer vision, and in consideration of the inherent characteristics of texture, that is, color and structure, an applicable approach for intelligent defect detection is proposed in this paper. The image of yarn-dyed fabric enhanced by fractional differentials is first converted from RGB true color space to L*a*b* color space, and energy-based feature images are acquired after the Log-Gabor filter filters chromatic and brightness channels. Then the paper defines the relations between energy and the local binary pattern as a new concept called energy-based local binary patterns (ELBPs). Finally defects can be detected, using ELBPs rather than grayscale-based local binary patterns. The proposed method can detect chromatic and structural defects. Experimental results for the defect detection from several collections of yarn-dyed fabrics indicate that a detection success rate of more than 94.09% is achieved for the proposed method, and the speed of test is also fast.