Application of Hausdorff distance in image matching

Abstract

In this thesis, a novel algorithm for image matching based points is proposed. The algorithm matches feature points extracted from the detected image with their corresponding points in the reference image. The fundamental problem of point matching is to get the affine transformation between two point sets, which transforms one point set so that its distance from the other point set is minimized. The algorithm procedure is optimized which adopts Similarity measurement method based on information metrics. Metric distance used in the algorithm is the Hausdorff distance. This algorithm has strong anti disturbance and fault tolerance. At last,the thesis also pointed out the direction of further study.