Saved in:
Bibliographic Details
Main Authors: Ma, Yanlong, Joshi, Nakul S., Robison, Christa S., Osteen, Philip R., Lopez, Brett T.
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2512.24384
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915700918976512
author Ma, Yanlong
Joshi, Nakul S.
Robison, Christa S.
Osteen, Philip R.
Lopez, Brett T.
author_facet Ma, Yanlong
Joshi, Nakul S.
Robison, Christa S.
Osteen, Philip R.
Lopez, Brett T.
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2512_24384
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Geometric Multi-Session Map Merging with Learned Local Descriptors
Ma, Yanlong
Joshi, Nakul S.
Robison, Christa S.
Osteen, Philip R.
Lopez, Brett T.
Robotics
Computer Vision and Pattern Recognition
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.
title Geometric Multi-Session Map Merging with Learned Local Descriptors
topic Robotics
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2512.24384