Ground Target Localization Algorithm Based on Unmanned Aerial Vehicle Image Analysis

Locating ground targets in unmanned aerial vehicle images is a key problem in computer vision and UAV applications. In this paper, we propose a novel ground target localization algorithm based on the image process technique. The main idea of this paper is to utilize scale-invariant feature transform (SIFT) feature descriptor to tackle the proposed problem. SIFT feature descriptor can extract feature points which are invariant to scaling, orientation, affine transforms and illumination changes. Firstly, SIFT descriptors are matched from UAV images and coarse positioning result, and location points are extracted from UAV images. Secondly, coordinates in coarse positioning results are gained from remote sensing images using the radiation transformation model, and the final ground target localization results are obtained from the coordinate transformation relation. Experimental results demonstrate that the proposed algorithm can detect and locate ground target in UAV images with high accuracy.