A Rendering Method of Laser Scanned Point Clouds of Large Scale Environments by Adaptive Graphic Primitive Selection

In this paper, a rendering method of the laser scanned point clouds of large scale environments is proposed for supporting an easy and intuitive understanding of the scanned environments. In this method, an adaptive primitives selection model and hierarchical point representation are used in the rendering of the scanned environment. Local geometry of the objects are estimated by principal component analysis, and the graphic primitives for points are adaptively created for effective rendering. View-dependent LOD using point hierarchy and an adaptive primitives selection model are also achieved for efficient rendering. Some rendering results for point clouds acquired from different scanning systems are shown and compared with other methods.