부산대학교 도서관

Unmanned aerial vehicle (UAV)-based 3-D LiDAR mapping has received increasing interest from both the research community and industrial operations for the 3-D reconstruction of large and dangerous environments. When a light detection and ranging (LiDAR) sensor is rigidly mounted to a UAV, one usually orients it, such that the LiDAR’s upward direction coincides with the UAV’s flight direction. For a spinning LiDAR, this fully specifies the mounting orientation because of circular symmetry. However, we argue, in this article, that this configuration is suboptimal, especially when the LiDAR’s full horizontal field of view is not important, either because of a limited field of view of other sensors in case of sensor fusion or because the scene to (swath) map via LiDAR is less wide than the LiDAR’s field of view. Indeed, in this situation, many of the points the LiDAR samples are not of interest. Additionally, in the usual configuration, the UAV’s movement can compensate for the limited elevation resolution of the LiDAR, but much less so for the azimuth resolution. For a simplified theoretical model, we introduce a quality metric for mounting angles. Using a simulation, we then derive the optimal mounting angle, which we validate via real world flights. We suggest to rotate this standard mounting orientation over an angle of −84.5° over the UAV’s vertical direction.