Robust loop closure detection and relocalization with semantic-line graph matching constraints in indoor environments

• DOI: 10.1016/j.jag.2024.103844
• Publication Date: 2024-04-18T16:24:46Z
• Authors: Xiqi Wang; Shunyi Zheng; Xiaohu Lin; Qiyuan Zhang; Xiaojian Liu
• Keywords: Closure (psychology)

Loop closure detection (LCD) plays an essential role in the Simultaneous Localization and Mapping (SLAM) process, effectively reducing cumulative trajectory errors. However, conventional LCD methods often encounter challenges when dealing with variations illumination, changes viewpoint, environments weak textures. This is due to their reliance on low-level geometric or image features. To address these issues, we propose a robust method named SL-LCD, which integrates semantic information line features fully leverage content attributes within indoor scenes, thereby establishing reliable feature correspondence between query images loop images. For retrieval of candidate closed-loop images, construct semantic-line-segment topological graph introduce matching algorithm perform task. approach exploits spatial achieve complex scenes. Furthermore, present voxel-based generalized ICP (SVGICP) relocalization tailored for challenging enhancing accuracy such scenarios. Experimental results demonstrate that SL-LCD proposed this paper surpasses state-of-the-art methods, accurately detecting closed loops, eliminating drift.