Extraction of normalized Digital Surface Model (nDSM) from LiDAR Data in Forest Inventory Mapping

Abstract

Listen

Translate article

Abstract The retrieval of Light Detection and Ranging (LiDAR) data is a complex procedure that necessitates extensive processing in order to develop terrain and surface models and forest structure applications. The gradual acquisition of LiDAR information is required to create Digital Elevation Models (DEM) and Digital Surface Models (DSM). The purpose of the study was to generate topographic DEM and normalized DSM (nDSM) data from LiDAR point cloud and to outline the canopy height extraction procedure in the New Forest region of the United Kingdom. Later, under 21 random enclosures, a demonstration of how the nDSM can be used in forest inventory mapping was discussed. The results show that, of the various interpolation techniques used to generate DEM, IDW had the lowest RMSE value of 0.382. The Digital Terrain Model (DTM) was created using two neighborhood settings (3×3) and (30×30), with the last one showing higher accuracy. In the comparison of different interpolation techniques, Inverse Distance Weighting (IDW) was found to have the lowest RMSE value of 0.382. Finally, within the enclosures, the percentage of no trees (mostly shrubs), canopy height ranged 2-10m, 10-15m, and > 15 was mapped. Each enclosure with 40% of its area covered by trees taller than 15 m was assumed to be harvestable. The study demonstrated detailed algorithm-based LiDAR data extraction and processing, which can be used to explore and forecast terrestrial ecosystems with advanced longitudinal orientation potentialities.