Fuzzy modeling of natural terrain elevation from a 3D scanner point cloud

This paper reports on simplification of 3D point clouds for terrain maps. This is an important issue for mobile robot navigation and path planning in natural environments. Points cloud maps obtained from onboard 3D range-finders can provide a large amount of data which is not uniformly distributed and that can be incomplete due to occluded and out-of-range areas. This work exploits the suitability of fuzzy systems to manage uncertain and incomplete information in order to obtain a simplified representation of natural terrain elevation maps. This solution has been implemented with Matlab ANFIS. The method is tested with experimental data from a 3D laser scanner in an outdoor environment.