ClusterSLAM: A SLAM backend for simultaneous rigid body clustering and motion estimation

Computational Visual Media

Sheng Yang¹  Zishuo Zhao²  Jiahui Huang²  Shi-Min Hu²  Yu-Kun Lai³ 

[1] Alibaba A.I. Labs, 311121, Hangzhou, China;BNRist, Department of Computer Science and Technology, Tsinghua University, 100084, Beijing, China;School of Computer Science and Informatics, Cardiff University, CF24 3AA, Cardiff, UK;

关键词: dynamic SLAM;    motion segmentation;    scene perception;   

DOI  :  10.1007/s41095-020-0195-3

来源: Springer

【 摘 要 】

We present a practical backend for stereo visual SLAM which can simultaneously discover individual rigid bodies and compute their motions in dynamic environments. While recent factor graph based state optimization algorithms have shown their ability to robustly solve SLAM problems by treating dynamic objects as outliers, their dynamic motions are rarely considered. In this paper, we exploit the consensus of 3D motions for landmarks extracted from the same rigid body for clustering, and to identify static and dynamic objects in a unified manner. Specifically, our algorithm builds a noise-aware motion affinity matrix from landmarks, and uses agglomerative clustering to distinguish rigid bodies. Using decoupled factor graph optimization to revise their shapes and trajectories, we obtain an iterative scheme to update both cluster assignments and motion estimation reciprocally. Evaluations on both synthetic scenes and KITTI demonstrate the capability of our approach, and further experiments considering online efficiency also show the effectiveness of our method for simultaneously tracking ego-motion and multiple objects.

【 授权许可】

CC BY