A method of deriving features of building from LIDAR point clouds in urban area

This research paper aims at extracting features, especially the plane feature, of building from Light Detection And Ranging (LiDAR) point clouds in Urban Area, and with these features and information to build the model of object. Unlike modeling object in other fields, such as reverse engineering, surfaces of building usually consist of abundant big and plane surfaces which are significant features. In urban area, most buildings can be simplified to the models that are made up of approximate plane surfaces to which features of buildings refer, in this paper. Among these surfaces, there are distinct points and intersection lines (the edges and vertices of the building object). Plane detection, surface adjacency relations restoration, model parameter calculation, and model reconstruction constitute the main research of extracting geometric features and modeling the object. In this contribution, some examples of deriving information from the point clouds are presented to demonstrate the method. And, the results prove that when the geometric information of vertices, edges and normal lines of the plane surface, together with the topological relations (Surface adjacency relations) among them, is derived from the point cloud data, the model, representing building in this paper, can be built effectively. Equally significant, models generated from this method, while occupying less memory space, can store more comprehensive structural information, and have a better exhibitive effect. And this method can provide three dimensional data of buildings from point cloud data in applications like GIS, navigation and virtual city.