Impacts of LiDAR Sampling Methods and Point Spacing Density on DEM Generation

Abstract

ABSTRACTLight Detection and Ranging (LiDAR) technology is gradually being adopted as the primary technique for surface structure derivation. There is a tendency for users of LiDAR data to want the highest density data possible, but such dense data sets increase overall costs including financial, processing, and storage without significant value added to the derived outputs, particularly in the case of digital elevation models (DEMs). In this study, we employed three sampling methods (systematic sampling, simple random sampling, and stratified random sampling) to select points from the original LiDAR data set that represented six different sparse densities. Using the sampled data sets, DEMs were generated and sensitivity analyses were conducted to explore whether an optimum LiDAR point sampling method and density could be identified. Despite the fact that significant differences in descriptive statistics were not observed for the three sampling strategies or point densities, there are evident differences among the derived DEMs, variograms, and viewsheds for very sparse samples (e.g., less than 2 percent of the original data set). In particular, the systematic and random sampling methods resulted in noticeable differences compared to the stratified random sampling approach. This research contributes to a better understanding of how LiDAR point sampling can be used to reduce data volume without compromising the final DEM quality.