A Hybrid Method Combining Hausdorff Distance, Genetic Algorithm and Simulated Annealing Algorithm for Image Matching

Object matching is an important issue for machine vision, object recognition and image analysis. A Hausdorff distance is one of commonly used measures for object matching because it is simple and insensitive to changes of image characteristics. But the conventional Hausdorff distances require high computational complexity and are not suited to the practical applications. In this work, we propose a new algorithm combining an improved partial Hausdorff distance, genetic algorithm and simulated annealing algorithm for higher computing efficiency and better matching results than the conventional Hausdorff measures. In this proposed method, the GA provides a global search and the SA algorithm provides local search. Analysis found the hybrid GA-SA conducts parallel analyses that increase the probability of finding an optimal solution while reducing computation time for object matching. Theoretical analysis and simulation results show that the new algorithm is very effective and robust under several kinds of noise conditions.