Optimal Transformation Parameter Optimization with Genetic Algorithm in Image Registration Within Hausdorff Distance
Institute of Information Technology, Zhejiang Shuren University, Hangzhou 310015, Zhejiang, China.
AbstractAs for the sensitivities of traditional Hausdorff distance to the noise and isolated point, which contribute to the lower matching ratio, this paper puts forward an improved Hausdorff distance model by genetic algorithm to optimize the transformation parameters. On the basis of a comprehensive analysis of the theory frame from different images matching techniques, a combined algorithm idea is proposed, using Hausdorff distance as the image measure function and using genetic algorithm as the search strategy to realize the image registration. Comparison with some recent patents on traditional algorithm，experiment shows that the improved Hausdorff distance by genetic algorithm can be a very good solution to robustness problem of the traditional algorithm, and has a higher matching speed in the case of the same edge points of image.
Genetic algorithm, grayscale, hausdorff distance, image registration, isolated point.
Download Free Order Reprints Order Eprints Rights and Permissions