FMD Stereo SLAM: Fusing MVG and Direct Formulation Towards Accurate and Fast Stereo SLAM
FMD Stereo SLAM: MVG and direct integration methods to achieve accurate and fast eyes SLAM
Fulin Tang, Heping Li, Yihong Wu
We propose a novel stereo visual SLAM framework considering both accuracy and speed at the same time. The framework makes full use of the advantages of key-feature based multiple view geometry(MVG)and direct-based formulation. At the front-end, our system performs direct formulation and constant motion model to predict a robust initial pose, reprojects local map to find 3D-2D correspondence and finally refines pose by the reprojection error minimization. This frontend process makes our system faster. At the back-end, MVG is used to estimate 3D structure. When a new keyframe is inserted, new mappoints are generated by triangulating. In order to improve the accuracy of the proposed system, bad mappoints are removed and a global map is kept by bundle adjustment. Especially, the stereo constraint is performed to optimize the map. This back-end process makes our system more accurate. Experimental evaluation on EuRoC dataset shows that the proposed algorithm can run at more than 100 frames per second on a consumer computer while achieving highly competitive accuracy.
We propose a novel binocular vision SLAM framework, taking into account the accuracy and speed. The framework based on key features of use of the multi-view geometry (MVG) and methods based on direct advantage. In the front end, we have performed a direct system and constant motion model equations to predict steady initial position, re-projected local map to find a corresponding 3D-2D, and by re-projection error minimizing finalized posture. The front-end process to make our system more quickly. At the rear end, MVG for estimating a 3D structure. When you insert a new keyframe, generate new map points by triangulation. In order to improve the accuracy of the proposed system, delete the bad points of the map and keep the global map through bundle adjustment. In particular, in order to optimize the implementation of three-dimensional map constraints. This back-end processes to make our system more accurate. Experimental evaluation of EuRoC data sets show that the proposed algorithm can run at more than 100 frames per second on the consumer computer, while achieving accuracy highly competitive.