Visual analysis of large-scale LiDAR point clouds

In this study we analyzed a series of LiDAR point clouds acquired over Taijiang district (part of Fujian province, China). The objective was to detect and extract water surface area from individual LiDAR point cloud, in a parallel means. To this end, interactive visualization of fine-grained data, global cluster algorithms, and statistical investigation were applied. We first rasterized point clouds into a series of matrices; then a density based algorithm (DBSCAN) allowed us to isolate the single cluster features which represented the water surfaces. Base on the clustered features, our framework created polygons for further quantitative assessments. The computations for the cluster analyses and their parallelized versions were carried out using R free software for statistical computing and graphics.