Hausdorff Distance Image Registration based on Features of Harris and SIFT

摘要

Harris detection and SIFT(Scale-invariant feature transform) are frequently used to image registration, but they have the disadvantages of not adapting the variety of scale and low efficiency, respectively. Hausdorff distance method based on the feature of Harris and SIFT is discussed here. Corners of the reference and the registered image are extracted and fused by using Harris corner detection and SIFT feature retrieval to expend the search scope of corner. Then the similarity matching principle is adopted to get rid of the error corners and the improved Hausdorff distance algorithm is used to realize image registration. The experimental results demonstrate that the computation time declines about 45% compared with traditional Hausdorff distance algorithm. The method has stronger anti-noise ability and rotary robustness, which improves the efficiency and accuracy of registration.