A UAV Visual SLAM System for Post-disaster Emergency Mapping

Post-disaster emergency surveying and mapping information is a key information resource to assist post-disaster rescue and reconstruction. In recent years, visual simultaneous localization and mapping (SLAM) technology has been expanding its application scenarios by virtue of its advantages of simple structure and good mapping effect. In this paper, an unmanned aerial vehicle (UAV) visual SLAM system for post-disaster emergency mapping is presented. Firstly, this paper analyzes the current situation of UAV mapping technology and visual SLAM technology. Subsequently, the hardware architecture and algorithm of the system are introduced. In terms of algorithm, the system first completes the construction of sparse feature point map based on ORB-SLAM2 algorithm, and then improves the algorithm through map reconstruction technology to realize the construction of dense point cloud map. After completing the system design, the whole system is applied to the real scene for experimental verification. Experimental results show that the designed system can realize the construction of a complete global graph of the real scene, and give the mapping task real-time, safe and efficient advantages.