Visual Point-Based Analysis of Laser-Scanned Historical Structures

Recently, terrestrial laser scanners (TLS) are actively used to acquire point cloud data of historical sites, and visualization of these data is becoming an important research target. A major purpose of the visualization is to extract 3D structures of laser-scanned historical sites such as culturally important buildings and monuments. In our previous work, we reported the stochastic point-based rendering (SPBR), which we recently proposed, realizes quick and precise transparent visualization of laser-scanned point cloud data. In this paper, we further extend this work by proposing a new method for visual analyses of laser-scanned point cloud data. We analyze point density distribution by projecting 3D points onto a plane and executing a logical operation process with the analyzed result. This prescription can extract information that is useful for porting laser-scanned point cloud data to a CAD system. For example, positions of wall, floor and ceiling can be automatically extracted.