Abstract: In order to solve the problem of low classification rate of illumination texture color image classification, this paper uses a color local angle binary patterns method for color texture feature extraction and presentation. On one hand, the method solves the problem of LBP for lacking description to natural texture image spatial structures. On the other hand, it improves the classification rate of illumination texture image. In order to verify the validity of the proposed method, the paper first extracts CLABP features from Outex texture dataset, and then classifies texture image by using nearest neighbor matching. The experimental results demonstrate the classification rate of this method achieves 90%.

Key words: illumination invariance property, local binary pattern, color local angle binary pattern, texture classification