Study On The Classification Method Of Animal Leather Based On Texture Feature

China owns a large leather industry and huge leather specimens need to be inspected during design,manufacture and marketing which mainly relies on the sense organ appraisal from human perception and is difficult to meet the demanding of the consistency and stability. In the process of leather identification, the artificial visual inspection influenced by subjective factors which due to inaccuracy. Besides, chemical method and biological method are complex and difficult to meet the demanding of social development. In this paper, digital image processing technology was utilized to fulfill feature extraction and classification based on leather texture,includes:(1) Image acquisition and preprocessing. Images were processed as gray image,the illumination was uniformed and noise was preliminary eliminated after leather image were taken by digital camera;(2) Texture feature reconstruction. Tiny texture of the image was removed using variation model main texture was retained. New texture feature map is reconstructed by using distance transform and skeleton extraction algorithm;(3) Texture feature discrimination with LBP-V. After completing the image texture features,the LBP-V method is used to extract the texture images. Besides, two types of library were set up: one is the texture training set, the other is the test set, and the histogram of the LBP-V map is used to determine the classification of the test image.The booming development of the leather industry and the huge market demand require the experimental personnel to effectively and quickly identify the different types of leather. Therefore,it is very important to study the identification technology of the advanced non- contact non-contact identification technology and realize the quick identification of leather.