LAUSR

During the last decades, with the increasing use of unmanned aerial vehicles (UAVs), applications that require automatic processing have been growing. One relevant process is the automatic estimation of digital elevation models (DEMs) for specific areas, for example, to support autonomous navigation. In many cases, autonomous flights are desired. Successful autonomous aerial navigation is dependent on several factors, such as UAV position information. In this case, some methods can be employed, combining Global Navigation Satellite System (GNSS) with Inertial Navigation Systems (INSs). When no GNSS and INS signals are available, computer vision has been an alternative for the UAV position estimation. This letter presents a novel approach to automatically estimate DEM from aerial images. Its major contribution is the use of characteristic points, the Zernike Moment Algorithm (ZMA), and the particle collision algorithm to find the correspondences between aerial images.