基于ORB-SLAM2的三维占据网格地图的实时构建

针对ORB-SLAM2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问题,提出了一种基于ORB-SLAM2的跳表地图(Skip List Map)构建算法,可用于三维占据网格地图实时构建。首先搭建了一个用于三维占据网格地图实时构建的Skip List Map模型,其时间复杂度仅为O(lgn);其次对Skip List Map三维占据网格地图的生成与更新做了详细推导;最后设计了ORB-SLAM2与Skip List Map算法相结合的方案。通过效率对比实验,表明本文算法具有较高的时间效率与灵活性;搭建实验所需平台并进行了真实场景实时实验,实验表明本文算法能实现三维网格地图的实时构建;且能清晰标识出环境中障碍物的位置,验证了本文算法的有效性。

Real-time Construction of 3D Occupied Grid Map Based on ORB-SLAM2

Considering the ORB-SLAM2 system can only output the motion trajectory map of the camera but the maps for path planning and navigation cannot generated, an algorithm based on ORB-SLAM2 for SkipList Map is proposed, which can be used for three-dimensional occupation. Firstly, a SkipList Map model for real-time construction of 3D occupied grid maps is built, and its time complexity is only . Secondly, the generation and update of SkipList Map 3D occupied grid maps are deduced in detail. Finally, a method combing ORB-SLAM2 and SkipList Map algorithm is designed. The efficiency comparison experiments show that the proposed algorithm has high time efficiency and flexibility. The platform needed for the experiment is built and the real-time experiment of the real scene is carried out. The experiment shows that the algorithm can realize the real-time construction of the 3D grid map and clearly identify the location of obstacles in the environment, which verifies the effectiveness of the proposed algorithm.