RAPID CHARACTERIZATION OF DENSE MATCHING SUITABLE FOR UAV VIDEO IMAGES

Image matching is between two or more images find matching points of the same name. Image dense matching is an important guarantee for three-dimensional reconstruction, extraction precision DEM, DSM. Since the UAV has flexible features, and can work in a complex environment, which makes UAV used is widely. UAVs can obtain a high degree overlapping images, which has a very important role for urban reconstruction and the map data update. UAV video image data acquired high degree of overlap is very large, feature extraction and feature generator consumes much time, how to effectively reduce the feature extraction and feature generates the amount of time? This paper presents a high degree of overlap for UAV video imaging characterization methods. The method uses Harris corner detection operator, then adopts feature descriptor simplified-DASIY (abbreviation: S-DASIY) to characterize detecting corners and generate the 25-dimensional feature descriptor for the corners; and in accordance with the appropriate matching criteria to match the feature points of the images, to get the match points between images. Characterized by experiments herein described method can effectively reduce the amount of time characterization.EICPCI-S(ISTP)zhaohy@pku.edu.cn810-8132015-Novembe