大尺度弱纹理场景下多源信息融合SLAM算法

In order to obtain the local accurate and global drift-free state estimation of an autonomous robot in the large-scale weak texture scenes, a SLAM system based on fusion of visual-inertial and global navigation satellite system (GNSS) is proposed. Firstly, by adding the line features to the local state estimation to represent the geometric structure information of the environment, the accuracy of the relative pose estimation between key frames in the weak texture scene is effectively improved. Secondly, by introducing a linear error representation, the linear feature is represented as a linear constraint on the end of the line, so the line feature is integrated into the linear representation based on the feature point algorithm, which effectively improves the robustness of the algorithm in the scene of the repeated line features. Finally, the multi-source information fusion algorithm is used to fuse the visual inertial and GNSS measurement information to achieve the local accurate and global drift free pose estimation, which effectively solves the problem of accurate state estimation in the large-scale weak texture scene. The evaluation results of several common datasets show that the proposed algorithm has stronger robustness and higher positioning accuracy.