SIMULATING UNMANNED-AERIAL-VEHICLE BASED LASER SCANNING DATA FOR EFFICIENT MISSION PLANNING IN COMPLEX TERRAIN

Abstract

In complex mountainous terrain the mapping efficiency is a crucial factor. Unmanned aerial vehicle (UAV) based laser scanning (ULS) has the capability for efficient mapping, as it allows realizing higher flight velocities, higher flying altitude above ground level (AGL) and larger distances between neighbouring flight strips, compared to image based techniques. However, fully utilising the efficiency of the system in mission planning (especially for complex terrain projects, where occlusions and differently inclined surfaces are present) is prone to miss the project requirements in terms of point density and strip overlap. Therefore, the numerical simulation of point densities is a helpful tool for realizing a reliable planning of scan coverage. We implemented a ray-tracing-based ULS-simulator, specifically designed for emulating the mechanism of a Riegl VUX-1LR laser scanner carried by a Riegl RiCOPTER. The simulator can consider copter and scanner motion, which makes it possible to generate synthetic scan data excluding or including the aircraft movement due to aerodynamics by using either planned trajectories from a flight planning software or recorded and post-processed trajectories from an inertial measurement unit (IMU). Laser shots are simulated by intersecting rays from the virtual scanner with a mesh-based digital surface model (DSM). The results show that the tool generates plausible synthetic laser point distributions. However, this is only the case, when aircraft aerodynamics are considered, as the effect of striping due to flight control corrections during the flight is very prominent. It can be shown that applying the presented tool for mission planning (without knowing the actual flight movements) has to consider an error margin of ±50pts/m2 in order to guarantee a compliance with the planned project requirements. Nevertheless, the consideration of terrain by a high resolution DSM, especially in complex terrain, improves the correlation between simulated and real point densities significantly.