Fast Algorithm for Image Registration Based on Hausdorff Distance

YANG Tong-yu,PENG Guo-hua
DOI: https://doi.org/10.3969/j.issn.1000-3428.2011.12.065
2011-01-01
Abstract:In the image registration process,the computational complexity is the major problem of the traditional Hausdorff distance algorithm. This paper proposes an image registration fast algorithm based on Hausdorff distance.Edge examination to the reference image and the image is registration,and then chooses a template,sets a variable threshold to improve the Hausdorff distance algorithm so as to reduce the amount of unnecessary to achieve rapid match.According to the matched data,it carries on the scale transformation and the rotation operation to the image, enables two images registration in the space.Experimental result proves that,comparing with the traditional matching algorithm,the computational complexity reduced greatly.
What problem does this paper attempt to address?