An Algorithm About 3D Buildings Extration From LIDAR Data Based On The Image

With the development of cyber city and the enhancement of urban planning level, it is very important to create 3D building model based on the prompt and accurate extraction of 3D building data. However, the traditional photogrammetry satisfies this requirement hardly. The Light Detection and Ranging (LIDAR) technology that appeared in recently years will offer a new way to resolve this problem. Because the characteristic of original LIDAR cloud data, such as form, density and spacing, in which 3D buildings are difficulty extracted purely. Many researchers have been attempting to study in this domain and developed several approaches for extracting building from LIDAR cloud data. This paper proposes a new practical algorithm for 3D building extraction from airborne LIDAR data.Firstly, the fundamental, components and application of LIDAR system are introduced respectively in this paper. Secondly, the original LIDAR data and digital image of the area are taken as experimental data and then the 2D buildings are extracted by processing the image. And the original LIDAR data has to be interpolated onto regular grid for data processing later. Thirdly, candidate regions for 3D buildings in LIDAR data are detected by matching the external image and LIDAR data. Finally, these initial building regions can be improved by the technique such as gradient threshold and all 3D buildings may be extracted precisely later.The new algorithm of 3D buildings extracted from LIDAR data based on the image is not affected by the terrain and building's shape. It will extract all 3D buildings in a wide range at low cost. Therefore, it has a wide range of application. The experimental results show the efficiency and feasibility of the new algorithm.