Combined Point and Line Features PL-SLAM Algorithm Research

Abstract

ORB SLAM is the current state of art SLAM algorithm both on accuracy and efficiency, includes map reuse, loop closing and relocalization capabilities. Owing to ORB SLAM is constructed based on point features, it works well in most cases, but loses tracking, or reduces its tracking accuracy in low-texture scenes or when the camera moves too fast causing lacking of reliable point features. To solve this problem, in this paper, we proposed a PL-SLAM (Point and Line based SLAM) algorithm which handle both points and lines to improve the tracking accuracy and reduce tracking failures. The PL-SLAM is an extension of ORB SLAM, focusing on how to introduce the line features into SLAM, establishing the optimization model that depends on point and line features, and solving it. Finally, the PL-SLAM algorithm is verified on the TUM RGBD benchmark, the results show that PL-SLAM not only improves the robustness in challenging environments, but also systematically improves the tracking accuracy in sequence frames.