LIDAR simulation and digital surface model processing

• المؤلف الرئيسي: Cheng, Lihang.
• مؤلفون آخرون: Tan Eng Leong
• التنسيق: Final Year Project
• اللغة: English
• منشور في: 2009
• الموضوعات: DRNTU::Engineering::Electrical and electronic engineering::Electronic systems::Signal processing
• الوصول للمادة أونلاين: http://hdl.handle.net/10356/17864
• المؤسسة: Nanyang Technological University

Object feature extraction and 3D model reconstruction methodology is desirable in many applications such as urban planning, data training, and simulations. The objective of this report is to reconstruct the 3D object models from Light Detection and Ranging (LiDAR) data by performing feature extraction on these objects. The general project procedure can be summarized as: 1) Prepare LiDAR data for visualization; 2) Object feature extraction based on object elevation values; 3) Comparison of the methods and discussion on the results. The main contributions to achieve these processes are the algorithms for LiDAR simulation. Firstly, Gridding algorithm rearranges the randomly scattered point data into orderly fashion, making the data convenient for further manipulation. Then the application of Normalized Digital Surface Model (NDSM) makes the classification of ground object based on elevations possible. Finally the Delaunay Triangulation Method, which joins the randomly distributed points together to form a continuous surface, provides a superior visualization of the LiDAR data. Other than the elevation information, object classification process also works based on the laser scanning properties and ground object surface properties.