Hierarchical Genetic Algorithm In Image Matching

With the development of image processing, image matching has been a part of a very considerable progress, which widely used in many other areas. How to guarantee the accuracy of matching and reduce the amount of calculation has been the key and difficulty in recent year. Research shows that, in all kinds of matching algorithms, the traditional matching algorithm can match a higher accuracy and it has a strong adaptability. In primitive conditions, the essence of process matching is that use the best matching criterion for the best search. Therefore, choosing an effective search criterion will greatly enhance the matching accuracy.In this work, a modified search strategy will be introduced during image matching process, which can achieve higher accuracy of image recognition and much better control experiments the best results, the main contents of this paper include:(1) Introduce the basic concepts of image matching technology, then image matching process will be transformed into a search process by describing its mathematical property; analyze the three key elements in the image matching technology:matching characteristics, similarity measure and search strategy, which result in the genetic algorithm search strategy;(2) Introduce the basic concepts of genetic algorithm and the algorithm implementation process, mainly focuses on genetic algorithm advantages, which including three key techniques of genetic. On this basis of the perspective of mathematical theory, prove the convergence of the algorithm in case of retaining the best individual strategy;(3) Comparing with the "point by point" matching algorithm in image template matching, propose a hierarchical genetic algorithm based on pseudo-parallel, design the rational genetic parameters used in image template matching. Compare the matching accuracy and matching time in different methods, while hierarchical genetic algorithm also has more certain research value and improvements.