Geometric Multi-Session Map Merging with Learned Local Descriptors
This work addresses the challenge of extended autonomous navigation by enabling consistent map alignment across sessions, though it appears incremental as it builds on existing learning-based descriptor and optimization methods.
The paper tackles the problem of merging point cloud maps from multiple sessions for autonomous operations in large-scale environments, achieving accurate and robust map merging with low error as demonstrated on public and self-collected datasets.
Multi-session map merging is crucial for extended autonomous operations in large-scale environments. In this paper, we present GMLD, a learning-based local descriptor framework for large-scale multi-session point cloud map merging that systematically aligns maps collected across different sessions with overlapping regions. The proposed framework employs a keypoint-aware encoder and a plane-based geometric transformer to extract discriminative features for loop closure detection and relative pose estimation. To further improve global consistency, we include inter-session scan matching cost factors in the factor-graph optimization stage. We evaluate our framework on the public datasets, as well as self-collected data from diverse environments. The results show accurate and robust map merging with low error, and the learned features deliver strong performance in both loop closure detection and relative pose estimation.