科研之友
微信
新浪微博
Facebook
分享链接
摘要
Modern urban environment is, unlike the past, composed of diverse features such as buildings, roads, vegetation and other structures and is fast changing day by day. Therefore, the study to comprehend and analyze such features in change has been carried out continuously in the field of remote sensing. In the past when the urban environment features were simple, the change of urban environment could be analyzed just through the satellite images of 2-dimensions. But, as the urban environment gets more complex and diverse, the necessity to analyze it with 3D information is growing more and more. The airborne light detection and ranging (LiDAR) is the best tool to analyze the urban environment and the studies related with it have been made actively in the remote sensing field. To extract diverse urban environment features through LiDAR, the prerequisite condition is the precise and correct generation of Digital Terrain Model (DTM).
DTM can be generated through the extraction of ground points from the airborne LiDAR data and then the interpolation. The ground points are usually extracted in the leveling method. If there is an elevated road in the target area, the elevated road is also extracted as a ground point to cause an error in generating DTM. Therefore, this study proposes an algorithm to detect the elevated roads which become the problem in extracting the ground points from airborne LiDAR data.
First, Digital Surface Model (DSM) is generated through the airborne LiDAR data; by applying mean planar filter, DSM is classified into planar surface and non-planar surface. Then, ground points are extracted by using the area-based filter. In the ground points extracted at this time, there are not only real ground points but also elevated road points. Therefore, to detect and eliminate the elevated road points among the extracted ground points, the regression plane is generated; the process is repeated to detect the elevated road points. In this process, diverse statistical figures are used to detect the points. Finally, DTM is generated with the ground points extracted through the above process.
To verify the validity of proposed method in this study, the method was applied to the airborne LiDAR of San Francisco area, USA and the quantitative precision was evaluated on the extraction of elevated roads. And, to see the influence of proposed method on the precision of DTM generation, the visual analysis on DTM without removal of elevated road points was additionally carried out. As a result, the proposed method turned out to have better performance in extracting the elevated road points and DTM generated in this method was more accurate too. Therefore, the technique proposed in this study seems useful in urban environment management, feature extraction and urban planning through DTM of urban environment.
- 出版日期2013-11
