Height gradient approach for occlusion detection in UAV imagery

Abstract

The use of Unmanned Aerial Vehicle (UAV) significantly increased in the last years. It is used for several different applications, such as mapping, publicity, security, natural disasters assistance, environmental monitoring, 3D building model generation, cadastral survey, etc. The imagery obtained by this kind of system has a great potential. To use these images in true orthophoto generation projects related to urban scenes or areas where buildings are present, it is important to consider the occlusion caused by surface height variation, platform attitude, and perspective projection. Occlusions in UAV imagery are usually larger than in conventional airborne dataset due to the low-altitude and excessive change in orientation due to the low-weight and wind effects during the flight mission. Therefore, this paper presents a method for occlusion detection together with some obtained results for images acquired by a UAV platform. The proposed method shows potential in occlusion detection and true orthophoto generation.