Tire pattern similarity detection based on template matching and LBP

Before testing the similarity of the tire pattern, the following problems must be solved: the background image of the tire is complicated, the angle of the pattern is varied, and the illumination is inconsistent. In this paper, the traditional template matching method is optimized by Hue, Saturation, Value color model (HSV) and Sobel operator, so that the acquired target image has higher accuracy. The original Local Binary Pattern(LBP) algorithm is improved to make it more suitable for tire pattern similarity detection while overcoming the problem of variable pattern angle.