| 摘要: |
| RGB-D SLAM是指使用RGBD相机作为视觉传感器，进行同时定位与地图构建（SLAM）的技术。RGB-D相机是近几年推出的能够同时采集环境RGB图像和深度图像的视觉传感器。首先对主流RGB-D相机，RGB-D SLAM算法框架流程做了介绍，然后对RGB-D SLAM算法的国内外主要标志性成果，以及RGB-D SLAM的研究现状进行介绍，并对RGB-D SLAM方法前端视觉里程计中特征检测与匹配、后端位姿图优化、回环检测等关键技术进行介绍总结。最后，对RGB-D SLAM算法的优缺点进行了分析，并对RGB-D SLAM算法的研究热点及发展趋势进行了讨论。 |
| 关键词: RGB-D相机 同时定位与地图构建 视觉里程计 位姿图优化 回环检测 |
| DOI： |
|
| 基金项目:国家国防基金（9140A09050313BQ01127）；国家自然科学基金（91120010） |
|
| An Overview of RGB-D SLAM |
| WANG Liu-jun,CHEN Jia-bin,YU Huan,ZHU Hui-shen |
| (School of Automation, Beijing Institute of Technology, Beijing 100081，China) |
| Abstract: |
| RGB-D SLAM refers to Simultaneous Localization and Mapping (SLAM) using RGB-D camera as a visual sensor. RGB-D camera is a kind of vision sensor which can be used to capture RGB images and depth images of environment. Firstly, this paper introduced the RGB-D camera used frequently and the RGB-D SLAM algorithm framework.Then the main achievements of RGB-D SLAM method at home and abroad, research status of RGB-D SLAM and the key technologies of the RGB-D SLAM method, such as feature detection and matching, the pose graph optimization of the back end and the loop closure detection were introduced and summarized. Finally, the advantages and disadvantages of the RGB-D SLAM method were analyzed, and the research hotspot and development trend of the RGB-D SLAM method were discussed. |
| Key words: RGB-D camera SLAM Visual odometry Pose graph optimization Loop closure detection |
