基于ORB-SLAM2的实时网格地图构建

针对目前视觉SLAM（同时定位与地图构建）系统只能输出相机的运动轨迹图而不能生成用于路径规划和导航地图的缺点，提出了一种基于ORB-SLAM2的网格地图实时构建算法。首先，建立了一个适用于视觉SLAM的逆传感器模型（inverse sensor model，ISM）；针对ISM模型重新编排了网格地图算法的构建机制，并对其进行详细推导；最后，介绍了ORB-SLAM2网格地图构建的具体实施方案。通过实验，对ISM模型和网格地图模型进行分析，确保了算法的可行性；用单目相机和RGB-D深度相机进行实时实验，实现了网格地图的实时构建，且能够清晰地显现出障碍物位置，验证了所提算法的有效性。

Real-time occupancy grid mapping using ORB-SLAM2

Currently the visual SLAM system can only output the camera''s motion trajectory, but it cannot generate maps for path planning and navigation. In order to solve this problem, this paper proposed a real-time grid map algorithm based on ORB-SLAM2. Firstly, it established an ISM for visual SLAM. Secondly, it rearranged the construction mechanism of the grid map algorithm for ISM and then derived it in detail. Finally, the paper introduced the specific implementation scheme of ORB-SLAM2 grid map construction. Through experiments, the algorithm shows its feasibility based on the analysis of the ISM and the grid map model. Furthermore, the real-time experiments using monocular camera and RGB-D camera can realize the real-time construction of the grid map and clearly show the positions of obstacles, which verifies the effectiveness of the algorithm.