Place recognition is a core component of Simultaneous Localization and Mapping (SLAM) algorithms. Particularly in visual SLAM systems, a robot must be able to recognize previously-visited places by measuring the appearance similarity between images representing these locations. However, it is sensitive to visual appearance change and also can be computationally expensive. In this paper, we propose an alternative approach that adapts LiDAR descriptors on 3D points obtained from stereo-visual odometry for place recognition. 3D points are more reliable than 2D visual cues (e.g., 2D features) against environmental changes (e.g., variable illumination) which benefits visual SLAM systems in long-term deployment scenarios. Stereo-visual odometry generates 3D points with a consistent scale, which enables us to use global LiDAR descriptors for place recognition with the goal of high computational efficiency. Through extensive evaluations on standard benchmark datasets, we demonstrate the accuracy, efficiency, and robustness of using 3D points for place recognition over 2D methods.
2. M2DP augmented with grayscale intensity
3. Scan Context augmented with grayscale intensity
Results on Oxford RobotCar dataset
Paper (IROS 2020):