Improved RANSAC UAV Image Matching Method in Low-visibility
In order to solve the problems of poor overhead image accuracy,reduced number of effective feature points and low robustness of UAV(Unmanned Aerial Vehicle)in low visibility environment,a low visibility UAV image matching method based on improved RANSAC(random sample consensus)is proposed.Firstly,the Daisy descriptor is added to the SURF(Speeded Up Ro-bust Feature)algorithm to improve the number of feature point extraction,and increases the robustness of the algorithm in low visi-bility environment.Secondly,the RANSAC algorithm is improved,and the block method and dynamic threshold are used to elimi-nate false matching of images and improve the matching accuracy of feature points.Finally,by using the VisDrone dataset,the ran-domly selected pictures are verified for feature point detection,and the traditional RANSAC algorithm is compared with the im-proved RANSAC algorithm to verify the proposed algorithm.The experimental results show that the feature point detection speed is greatly improved,and the matching accuracy of the improved RANSAC algorithm in different environments is improved by 9.5%and 16.9%compared with the traditional RANSAC algorithm.
万方数据
维普
