Leveraging Structural Information to Improve Point Line Visual-Inertial Odometry

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