Leveraging Structural Information to Improve Point Line Visual-Inertial Odometry

Leveraging line features to improve the accuracy of the SLAM system has been studied in many works. However, making full use of the characteristics of different line features (parallel, non-parallel) to improve the SLAM system is rarely mentioned. In this paper, we designed a VIO system based on points and straight lines, which divides straight lines into structural (that is, straight lines parallel to each other) and non-structural. To optimize the line features effectively, we used two-parameter representation methods for both structural and non-structural straight lines. Furthermore, we designed a stable line matching method based on frame-to-frame (2D-2D) and frame-to-map (2D-3D) strategies which can significantly improve the trajectory accuracy of the system. We conducted ablation experiments on synthetic data and public datasets, and also compared our method with state-of-the-art algorithms. The experiments verified the combination of different line features can improve the accuracy of the VIO system, and also demonstrated the effectiveness of our system.