Study On Image Matching Technology Combination With Optimization Algorithm

With the progress of science and technology and the development of an era,machine vision has gradually appeared in the line of people's sight.As the basis of visual computing,image matching technology is attract increasingly with people.At present,the improved algorithm of image matching emerge in endlessly.It is make the quality of performance with algorithm to make the result of the matching on the speed and accuracy.There are many advantage of image matching based on feature,such as real-time performance,the amount of information reflected by the feature,tiny time complexity and so on.But due to the result of image matching is barely satisfactory,which is affected by some external factors,for instance,shooting conditions,light intensity and geometric distortion,etc.So far,there have been some scholars to apply genetic and particle swarm intelligence algorithm in image matching,It is used to improve the performance of matching algorithm.The addition of optimization algorithm to improve image matching accuracy has been effective.In this paper,the main research content as follows:(1)Using the topology of the particle swarm algorithm was improved,this paper proposed a topology optimization hierarchically.This is based on tree structure that upward traversal hierarchically.It is greatly shorten the search time of population.(2)To improve the parameters in the formula of essential particle swarm,adding the topology optimization hierarchically,an improved particle swarm algorithm with the best leaf nodes has been proposed.The algorithm to break through the limitations of original particle swarm algorithm,getting full use of the diversity of particle in the population,thus make a more comprehensive search.(3)The improved particle swarm algorithm merged into the field of image matching,proposed a fusion particle swarm and the topological similarity of image matching method.In the matching process,firstly to detect the feature points,using the contour similarity calculate the obvious characteristics distribution in the area.Secondly using the topological similarity calculation the structure of the relationship and distribution of similarity degree between the image characteristics.Utilizing the improved particle swarm algorithm to optimize a characteristics within the contour scope.Finally making use of topology to simplify the constraint relations.This algorithm not only improve the accuracy of area,but also promote the speed.