COMPARISON OF FOREST STRUCTURE METRICS DERIVED FROM UAV LIDAR AND ALS DATA
- 1Department of Geodesy and Geoinformation, TU Wien, 1040 Vienna, Austria
- 2Remote Sensing Laboratories, Department of Geography, University of Zurich, 8057 Zurich, Switzerland
Keywords: Point height distribution, Fractional cover, Vertical complexity index, Voxel metrics, Occlusion effect, Acquisition characteristics
Abstract. Point clouds derived from airborne laser scanning (ALS) and from LiDAR sensors mounted on unmanned aerial vehicles (ULS) reveal differences caused by the different sensor systems and acquisition geometries. These differences in the system characteristics are reflected in forest structure metrics that are derived from the respective point clouds. In our study, we investigate the completeness of scene coverage between the two systems and address differences between structure metrics derived from ULS and ALS, namely in point height quantiles, fractional cover (fc), the vertical complexity index (VCI) and the number of canopy layers (nLayers). The metrics are evaluated for raster cell sizes of 1–10 m in order to investigate the spatial scale on which the sensor systems provide comparable metrics. We found highest correspondences between ALS and ULS in the VCI- and the nLayers-metrics, while fc revealed large differences. For the height quantiles, the absolute differences were larger for the 10%- (h10) and the 50%- (h50) than for the 90%- (h90) height quantile. Furthermore, we found differences between ALS- and ULS-metrics to decrease for larger cell sizes, except for fc, for which the differences increased, and h50 and h90, respectively, for which the differences were relatively stable for all cell sizes.