Real-time Meshing of Point Clouds (DAAV)

This thesis reviews state of the art techniques to reconstruct a surface out of a 3D point cloud. A new approach is implemented and evaluated based on an existing algorithm that triangulates certain well sampled regions and calculates connectivity information for sparse and noisy point sets. A noise model is introduced to find a manifold triangulation for those areas that cannot be interpolated by the existing method. As new data points are added to the point cloud, they are merged on the fly with the already reconstructed geometry.