Stereo Visual Inertial SLAM Algorithm Fusing Point and Line Features

Traditional Visual-Inertial SLAM algorithms often have poor localization accuracy or even failure in weak texture scenes. In this paper, a stereo visual inertial SLAM algorithm based on point and line features is proposed. The algorithm is based on the open source ORB-SLAM3, on which line features are added. Improved LSD is used to detect line features, IMU or constant velocity model is used to improve the efficiency of line feature projection matching, point and line reprojection error is constructed, and the position and pose are solved jointly IMU. Cauchy robust kernel function is used to reduce the influence of feature mismatching on the optimization process. The open source EuRoC data set is used to evaluate the proposed algorithm, and the experimental results show the effectiveness of the algorithm in this paper.