Image Matching Method Based On Image Enhancement Technology And SURF Algorithm

In order to make image change detection,image fusion and splicing quickly and efficiently,matching the collected images accurately is firstly needed.In the condition of the fog weather,the visibility of the sky is low,the obtained images from instruments give a bad visual to the human eyes.With the image analysis processing,fewer image information is obtained,using general image matching method to get high precision result of the matching.With the related basic theory of the image processing,this paper makes a deep exploration of the image enhancement,the structure of image feature extraction,the distance scale invariance of the similar triangular and relevant technology and basic theory of image matching.This paper has a detailed introduction to the interest point detection operator of Moravec?corner extraction operator of Harris?the feature detection operator of SUSAN based on gray scale and the feature detection algorithm of SIF based on scale invariant.SIFT algorithm keep invariance to the image of rotation changes,the image of size changes and the light intensity change is verified by simulation experiments,also has the very strong resistance to the noise interference from the outside world.The SITF algorithm using one hundred and twenty eight dimensions descriptor vector to describe the feature points.There is no doubt increase the computation of the similarity measure function between the feature points.And that there are many issue points in the features matched by SIFT algorithm.This paper combines image enhancement algorithm MSR with SURF matching algorithm for remote sensing image contrast research,proposed an image matching method based on image contrast enhancement of the SURF.First,using the MSR algorithm to enhance the image which obtained in the fog weather,then completed the feature detection with SURF algorithm,measure the similarity of feature points with Euclidean distance.Eliminate false matching points with the distance scale invariance of similar triangles.Simulation experiments show that the proposed method significantly improves the image contrast problem of poor contrast,enrich the image details,obtain a lot of rich information in the process of extracting feature points and matching results got obvious improvement.Deleting the point of low matching characters,further improved the precision of image feature points matching,provides a good foundation to the image change detection,fusion and splicing.