基于地图拼接的连续定位SLAM方法研究

Under the condition that there is a large-scale map, it is difficult to achieve high-precision positioning of the area outside the map through the ORB-SLAM2 system. The method proposed in this paper has improved the system to add the ability to create multiple maps. It enables the ORB-SLAM2 system to reconstruct a map and perform simultaneous positioning and mapping when it fails to track the large-scale map. It can achieve high-precision positioning in unfamiliar environments outside the saved map. In addition, it realizes the function of map switching and stitching, and can obtain continuous trajectories. The experimental results show that the proposed method improves the overall positioning accuracy and the success rate of positioning is 98.61%, thus increasing the environmental adaptability of ORB-SLAM2 system.