Speed-up Matching Method With Navigation Data for UAV Remote Sensing Images of Coastal Region

摘要

UAV remote sensing platform can obtain images of target regions quickly. It has distinct advantages in the detection of oil spill, the emergency response of searching and rescuing, the survey of coastal regions, etc. However, the existing UAV images are difficult to meet the needs of rapid processing, because the amount of their data is too large and the traditional processing methods take too much time in the matching stage. This paper designs a speed-up matching algorithm which utilizes navigation data in UAV to get the elements of exterior orientation. The algorithm is based on the collinearity equation with the flat terrain in the coastal regions. Those elements can be used to compute the original homography matrix and overlapping regions. After extracting interest points by SURF algorithm, the matching method only chooses some points in overlapping regions for matching stage. The algorithm can improve the matching speed, and also can decrease mismatching to improve the accuracy.