Passage-Aware Structural Mapping for RGB-D Visual SLAM
Ali Tourani, Miguel Fernandez-Cortizas, Saad Ejaz, David Pérez Saura, Asier Bikandi-Noya + 2 more
TLDR
This paper introduces a passage-aware RGB-D VSLAM system that detects and models doorways and passages for improved indoor robot navigation.
Key contributions
- Detects doors and traversable openings by fusing geometric, semantic, and topological cues.
- Models doors as planar entities within walls, classifying them as traversable or non-traversable.
- Infers passages using both camera-wall interaction evidence and geometric wall discontinuities.
- Integrates into vS-Graphs, enhancing scene graphs with passage-level abstractions and room connectivity.
Why it matters
This work addresses a critical gap in VSLAM by explicitly modeling doorways and passages, which are vital for indoor robot navigation. By enriching scene graphs with these structural elements, it significantly improves room connectivity and lays the groundwork for more intelligent, BIM-informed robotic systems.
Original Abstract
Doorways and passages are critical structural elements for indoor robot navigation, yet they remain underexplored in modern Visual SLAM (VSLAM) frameworks. This paper presents a passage-aware structural mapping approach for RGB-D VSLAM that detects doors and traversable openings by jointly fusing geometric, semantic, and topological cues. Doors are modeled as planar entities embedded within walls and classified as traversable or non-traversable based on their coplanarity with the supporting wall. Passages are inferred through two complementary strategies: traversal evidence accumulated from camera-wall interactions across consecutive keyframes, and geometric opening validation based on discontinuities in the mapped wall geometry. The proposed method is integrated into vS-Graphs as a proof of concept, enriching its scene graph with passage-level abstractions and improving room connectivity modeling. Qualitative evaluations on indoor office sequences demonstrate reliable doorway detection, and the framework lays the foundation for exploiting these elements in BIM-informed VSLAM. The source code is publicly available at https://github.com/snt-arg/visual_sgraphs/tree/doorway_integration.
📬 Weekly AI Paper Digest
Get the top 10 AI/ML arXiv papers from the week — summarized, scored, and delivered to your inbox every Monday.