Techniques for the Visualization of LiDAR Data

Point clouds are a popular and abundant data format in computer graphics. In particular, topographic and GEO data are typically generated using LiDAR technology and stored as point clouds. Therefore, LiDAR-generated data is often visualized directly as point clouds. Visualization of these point clouds without advanced techniques is less than ideal; data generated by airborne LiDAR is sparse, resulting in gaps between points; another potential problem is the difficulty in identifying and distinguishing different volumetric structures due to the sparsity. This thesis examines different visualization techniques for LiDAR point cloud data and looks at the advantages and disadvantages of these techniques. The implementation part of this thesis focuses on a voxel-based visualization that improves on the listed drawbacks of point-based rendering. The results of this implementation show that the voxel-based visualization, although it has a larger memory footprint, does a good job of solving the gap problem. Additionally, this work shows that rendering certain objects differently based on the object classification can improve the user experience.