Surround-View Semantic BEV SLAM: A Direct-Method Approach for Robust Mapping in Indoor Garages

View at publisher Summarise with VIRAT
Authors: Zihan Yuan, Junqiao Zhao
Cited by: 0
References: 12

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Autonomous Valet Parking (AVP) represents a critical application of autonomous driving technology. However, in texture-poor environments such as underground garages, traditional SLAM methods often struggle to maintain robustness and accuracy. To address this challenge, this paper proposes a Semantic BEV SLAM system that leverages semantic Bird's-Eye-View (BEV) projections. By integrating direct matching, featurebased loop closure detection, and global pose graph optimization, the proposed framework effectively overcomes the limitations of conventional approaches. Experimental results demonstrate that the system achieves centimeter-level mapping accuracy, effectively eliminating accumulated drift caused by long-distance driving while generating high-quality global semantic maps. With its high precision and strong robustness, the proposed system offers a reliable solution for AVP applications.

Fields of Study & Subjects

Robotics and Sensor-Based Localization Autonomous Vehicle Technology and Safety 3D Surveying and Cultural Heritage

Publication details

Journal / source
International Conference on Electrical Automation and Artificial Intelligence (ICEAAI)
Pages
1–4
Publication date
16 January 2026
Language
EN