Occlusion mapping reveals the impact of flight and sensing parameters on vertical forest structure exploration with cost-effective UAV based laser scanning

Recent studies have demonstrated the potential of light detection and ranging (LiDAR) from uncrewed aerial vehicles (UAVs) for assessing forest structures. Maximizing data completeness and representativeness is essential to accurately retrieve key structural parameters. However, knowledge on how data acquisition approaches affect canopy volume exploration is sparse. This study investigated the effects of selected sensing and flight settings on canopy occlusion in a central European forest using a cost-effective sensor system. We conducted 44 flights with a DJI Matrice 300 RTK UAV and a DJI Zenmuse L1 LiDAR sensor, with different combinations of flight speed, azimuthal flight directions, sensor tilt angles, and scan patterns. Using sensor position reconstruction and a ray tracing algorithm to quantify occlusion, we found that: (1) A larger sensor tilt angle up to 15° increased total occlusion, enhancing exploration in the upper canopy, while decreasing it below due to reduced canopy penetration. (2) Flying multiple azimuthal directions with a linear scanning mode reduced vertical occlusion by up to 15.0% due to improved coverage from diverse perspectives. (3) Lissajous scanning patterns resulted in 10.1% less vertical occlusion compared to linear patterns, underscoring the importance of additional viewing angles. Based on these findings, we recommend: (a) incorporating nadir sampling for below-canopy assessment; (b) using off-nadir angles for upper canopy evaluation; and (c) maximizing sampling perspectives and viewing angles to reduce occlusion effects. Our results offer transferable insights to optimize UAV LiDAR data acquisitions, thereby contributing to an enhanced structural metric retrieval and improved analysis of forest functional properties.