OpenMulti: Open-Vocabulary Instance-Level Multi-Agent Distributed Implicit Mapping

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Dou, Jianyu, Deng, Yinan, Wang, Jiahui, Tang, Xingsi, Yang, Yi, Yue, Yufeng
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916928056983552
author Dou, Jianyu
Deng, Yinan
Wang, Jiahui
Tang, Xingsi
Yang, Yi
Yue, Yufeng
author_facet Dou, Jianyu
Deng, Yinan
Wang, Jiahui
Tang, Xingsi
Yang, Yi
Yue, Yufeng
contents Multi-agent distributed collaborative mapping provides comprehensive and efficient representations for robots. However, existing approaches lack instance-level awareness and semantic understanding of environments, limiting their effectiveness for downstream applications. To address this issue, we propose OpenMulti, an open-vocabulary instance-level multi-agent distributed implicit mapping framework. Specifically, we introduce a Cross-Agent Instance Alignment module, which constructs an Instance Collaborative Graph to ensure consistent instance understanding across agents. To alleviate the degradation of mapping accuracy due to the blind-zone optimization trap, we leverage Cross Rendering Supervision to enhance distributed learning of the scene. Experimental results show that OpenMulti outperforms related algorithms in both fine-grained geometric accuracy and zero-shot semantic accuracy. In addition, OpenMulti supports instance-level retrieval tasks, delivering semantic annotations for downstream applications. The project website of OpenMulti is publicly available at https://openmulti666.github.io/.
format Preprint
id arxiv_https___arxiv_org_abs_2509_01228
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle OpenMulti: Open-Vocabulary Instance-Level Multi-Agent Distributed Implicit Mapping
Dou, Jianyu
Deng, Yinan
Wang, Jiahui
Tang, Xingsi
Yang, Yi
Yue, Yufeng
Robotics
Multi-agent distributed collaborative mapping provides comprehensive and efficient representations for robots. However, existing approaches lack instance-level awareness and semantic understanding of environments, limiting their effectiveness for downstream applications. To address this issue, we propose OpenMulti, an open-vocabulary instance-level multi-agent distributed implicit mapping framework. Specifically, we introduce a Cross-Agent Instance Alignment module, which constructs an Instance Collaborative Graph to ensure consistent instance understanding across agents. To alleviate the degradation of mapping accuracy due to the blind-zone optimization trap, we leverage Cross Rendering Supervision to enhance distributed learning of the scene. Experimental results show that OpenMulti outperforms related algorithms in both fine-grained geometric accuracy and zero-shot semantic accuracy. In addition, OpenMulti supports instance-level retrieval tasks, delivering semantic annotations for downstream applications. The project website of OpenMulti is publicly available at https://openmulti666.github.io/.
title OpenMulti: Open-Vocabulary Instance-Level Multi-Agent Distributed Implicit Mapping
topic Robotics
url https://arxiv.org/abs/2509.01228