The Evolution Of The Image Processing Method

Evolutionary computation is a generic method used to pursue the optimal solution in light ofthe hypothesis of evolution in the natural world. It mainly includes four branches, i.e. geneticalgorithms, evolution strategies, evolutionary programming and genetic programming. With itsintrinsic parallelism and intellectual traits such as self-organizing, self-adaptiveness andself-learning, evolutionary computation has successfully found its use in tackling knotty problemsthwarting conventional methods. Image matching is to seek commonnesses of two images with acertain degree of similarity for the purpose of image identification. In general, it is to search for atarget image similar to the source image through a search window. This paper aims at seeking grayvalue-based similarities, or rather, gray value-based template matching, between two differentimages, as the previous gray matching failed to produce a satisfactory result. The previous graymatching differs from the gray value-based template matching not only in manner of featureextraction, but in that it has some defects in seeking the optimal solution, such as enormouscalculation, time-consuming, poor matching, etc. In this paper, the author puts forward theapproach of evolution to image matching. After a detailed account of various types and methods ofmatching, pursuant to their respective characters, Guo Tao algorithm, as a type of evolutionarycomputation, is introduced. Based on experimentation, each evolutionary method is elaborated inrespect of its particular strong and weak points. Finally, a conclusion is reached that the problemsrelated to image processing can be addressed by means of evolutionary computation and withsatisfactory matching results.The paper is divided into three parts:First of all, it sets forth genetic algorithm and Guo Tao algorithm among others in respect oftheir principles and applications. Meanwhile, it reveals the advantages and disadvantages ofevolutionary computation including genetic algorithm, Guo Tao algorithm and artificial bee colonyalgorithm by procuring the maximum and minimum of a complex function.Secondly, a careful analysis is made on the image matching theory in image processing. Thetemplate matching is elaborated in its essence. The method of evolutionary search is introducedas well.Finally, by virtue of MATLAB, the good function of evolutionary computation includinggenetic algorithm, Guo Tao algorithm is verified.To sum up, this paper focuses on solving image matching from the perspective of evolutionarycomputation and on verifying its feasibility through experimentation. It is of reference value andsignificance to the application of evolutionary computation and to improve the effect of imagematching.