A Feature Points Matching Method for Calibration Target Images

Feature matching is a crucial research in computer vision. For realizing the effective and accurate feature matching for the feature points in the images of chessboard calibration target used in the wheel alignment system, the popular calculation algorithms are analyzed. The traditional polar constraint and polar equation for solving the problem constructing the relationship between corresponding points are indicated. A novel process which is special for matching calibration target images including sequential consistency constraint, polar constraint, distance constraint, uniqueness constraint are presented. Experimental results show that the matching precision and reliability of the method meet the requirements of the vision detection system for vehicle wheel alignment.