Graph-based SLAM (Simultaneous Localization And Mapping) for Bridge Inspection Using UAV (Unmanned Aerial Vehicle)

The manual bridge inspection requires the hard work of the surveyor. A robot such as UAV (Unmanned Aerial Vehicle) can be used to avoid boring and dangerous works and replace humans. The field of research for bridge inspection using UAV has gradually been developed to meet human needs. However, UAV's GPS (Global Positioning System) receiver cannot receive GPS signals under the bridge. This is because the satellite signal is blocked by the bridge structure. The purpose of this paper is to propose a localization method for bridge inspection using a UAV in the lower part of the bridge. Our localization method is a graph-based SLAM (Simultaneous Localization And Mapping) approach using a 3D LiDAR and a mono camera. VO (Visual Odometry) from the camera and the ICP (Iterative Closest Point) algorithm using a 3D LiDAR provide nodes and constraints for the graph structure. Experiments were conducted in a bridge environment and our method was compared with the ground truth obtained from an RTK (Real Time Kinematic) GPS.