Interpretation and Generalization of 3D Landscapes from LiDAR Data

Abstract

This paper discusses the interpretation and generalization of three-dimensional landscapes derived from laser scanning data. Airborne laser scanning systems provide an unparalleled detailed description of surfaces on a large scale. This level of detail, however, raises questions regarding the effective visual communication of the data. In this paper, we consider visualization from a cartographic perspective, particularly in terms of the generalization of details and enhancement of important features. Our goal is to visualize topography efficiently, while emphasizing such features as the road network. To achieve this, the communication of 3D landscapes has been implemented as a modular pipeline of individual, configurable processes, focusing on the interpretation of the laser dataset and its generalization. The raw laser point-cloud then becomes an input to a sequence of processes that result in a generalized topography. The interpretation process concerns the extraction of the terrain and the road network from the raw data. The generalization of the interpreted entities then follows. The application of the proposed model shows improved manifestation of the road network within the terrain.