Ground filtering and vegetation mapping using multi-return terrestrial laser scanning

摘要: Abstract Discriminating laser scanner data points belonging to ground from above-ground (vegetation or buildings) is a key issue in research. Methods for filtering into and non-ground classes have been widely studied mostly on datasets derived airborne scanners, less so terrestrial scanners. Recent developments sensors (longer ranges, faster acquisition multiple return echoes) has aroused greater interest surface modelling applications. The downside of TLS that typical dataset high variability point density, with evident side-effects processing methods CPU-time. In this work we use scan sensor which returns target echoes, case providing more than 70 million our study site. area presents low, medium vegetation, undergrowth varying as well bare morphology (i.e. very steep slopes flat areas). We test an integrated work-flow defining terrain model (DTM DSM) successively extracting information vegetation density height distribution such complex environment. Attention was given efficiency speed processing. method consists first step subsets the original define candidates by taking account ordinal number amplitude. A custom progressive morphological filter (opening operation) applied next, candidate using multidimensional grid fallout function distance scanner. Vegetation mapping over then estimated weighted ratio counts tri-dimensional space each cell. overall result pipeline clouds minimal user interaction, producing Digital Terrain Model (DTM), Surface (DSM), map Canopy Height (CHM). These products are importance many applications ranging forestry hydrology geomorphology.