An Improved Matching Algorithm in the Application of Robot Vision System

For 2D visual applications, using template matching algorithm to identify objects is difficult to get accurate matching results. Because the traditional template matching algorithm is based on gray-scale correlation, it is sensitive to rotation, illumination variation and noise. By improving the classical Canny operator, obtain the precise edge feature image. Employ the binarization to feature image to exclude the effect of variation in gray value caused by the illumination and noise. Then compensate the rotation angle of the target to the template and make template matching ultimately converted to binary match. Apply the algorithm proposed in this paper to 2D visual application and the experimental results prove that the algorithm can overcome the influence of illumination and rotation on the recognition process of the visual system.