Study Based On The Three-dimensional Reconstruction Of Vehicle Laser Scanning Data

The emergence of laser scanning technology for fast and accurate access to the 3D (three-dimensional) spatial information provides a new technology. The sensors of laser scanning system can be placed on different platforms, such as satellites, aircraft, trains, cars and other mobile carriers, so the methods of its data collection is flexible and adaptable. Vehicle-borne laser scanning system can obtain and update the 3D information of city objects, road traffic facilities, topography and other objects quickly, largely-scale and accurately, so it is very suitable for acquiring the 3D space information of urban quickly and accurately that 3D reconstruction on the basis of. But point cloud laser scanning system acquires is discrete points not conducive to express real scenes, so it needs the conversion from huge amounts of discrete points to 3D models and this process is commonly referred to 3D reconstruction. At present the method of 3D reconstruction based on vehicle-borne laser scanning is mainly dependent on human-computer interaction or semi-automatic processing, and it is not for the 3D of whole city objects. To solve this problem, this article only offers a method for the 3d reconstruction of point cloud from vehicle-borne laser scanning based on the classification of city objects. Finally using the point cloud data of a city based on the vehicle-borne laser scanning system that is self-developed of the capital normal university as experimental object, the result verifies the feasibility and practicability of the method that this article proposed.This paper is based on the vehicle-borne laser scanning system. Therefore, firstly it introduced the current vehicle-borne laser scanning system, and analysis the data it obtains. Then it briefly introduces the principles of the present 3D reconstruction, as well as 3D reconstruction methods based on airborne laser and ground laser scanning system, and analyzes the advantages and disadvantages of these methods to provide basis for 3D reconstruction of vehicle-borne laser scanning system. Finally, it's a detailed description of 3D reconstruction process based on vehicle-borne laser scanning system, with the pretreatment of raw point cloud data and 3D reconstruction based on classification. Firstly it describes how the raw data executes the data filtering, data fusion and data classification for the data suitable for 3D mode; then it focuses on how to process the different 3D reconstruction with the different categories, including the construction of Delaunay triangulation with the surface model such as ground,3D reconstruction with the body model such as houses and other construction of model with the point and line model. In addition it researches the problem of the buildings blocked by other features, and proposes a solution method using the feature that laser emission center point, occlusion point and blocked point are collinear. At last these functions are achieved on our project team's point cloud workstation platforms.Finally, this paper.summarizes the advantages and problems of this approach, prospects for this approach' application prospects and the research direction of data processing and 3D reconstruction with point clouds of vehicle-borne laser scanning.