Loop-Closure and Map-Combiner Detection Strategy based on LIDAR Reflectance and Elevation Maps

Loop-closure and map-combiner detection is very important to robustly enable applying SLAM technology and generate high-definition maps in level of cities and prefectures. In addition, the detection process paves the way to evaluate the map quality at the revisited areas in terms of relative and global position errors. Accordingly, this paper proposes a unique framework for precisely detecting loop-closure and map-combiner events regardless to the driving trajectory, the vehicle sensor configurations and the environment structure. A very challenging environment including long underground tunnel and multilayer junction has been selected to emphasize the reliability, scalability and robustness of the proposed framework. The experimental results have verified the capability to detect the map-combiner events precisely even-though with the existence of a considerable low positioning accuracy by a GNSS/INS-RTK system.