Image Matching Algorithm Based on an Improved Hausdorff Distance
Xiaohong Li, Yizhen Jia, Feng Wang, Yuan Chen · Advances in intelligent systems research/Advances in Intelligent Systems Research · 2013
As for the rotation, scaling, cropping in image matching, corresponding algorithm based on Hausdorff distance is deduced.The proposed algorithm extracts feature points by SIFT operator in original image and target image, and determines whether there is a kind of affine transformation between two images by an improved Hausdorff distance and box distance transformation.If there is an affine transformation which satisfies the requirement of the Hausdorff distance, the original image and target image can be matched.If there isn't an affine transformation which satisfies the requirement of the Hausdorff distance, the original image and target image couldn't be matched.The experimental results show that the algorithm can reduce affine space, improve the efficiency of image matching and eventually find the affine transformation between two images.