PeterFWS/Structure-PLP-SLAM

[ICRA'23] The official Implementation of "Structure PLP-SLAM: Efficient Sparse Mapping and Localization using Point, Line and Plane for Monocular, RGB-D and Stereo Cameras"

What it solves

Structure PLP-SLAM addresses the challenge of efficient sparse mapping and localization in various environments. By incorporating points, lines, and planes into the SLAM (Simultaneous Localization and Mapping) process, it improves the robustness and accuracy of camera tracking and environment reconstruction compared to systems that rely solely on point features.

How it works

The system processes visual data from monocular, RGB-D, or stereo cameras. It utilizes a combination of geometric primitives—points, lines, and planes—to build a sparse map of the environment. For planar SLAM, it integrates instance planar segmentation (using tools like PlaneRecNet) to identify planes in the scene. The system also supports re-localization, allowing a camera to find its position based on a pre-built map, and can optionally perform depth-based dense reconstruction for RGB-D data.

Who it’s for

This project is designed for robotics researchers and developers working on visual SLAM, specifically those needing high-performance localization and mapping across different camera configurations (monocular, stereo, and RGB-D).

Highlights

  • Multi-Primitive Mapping: Uses points, lines, and planes to create a more comprehensive sparse map.
  • Camera Versatility: Compatible with monocular, RGB-D, and stereo camera setups.
  • Re-localization Support: Ability to perform map-based image localization using pre-existing maps.
  • Integration with PlaneRecNet: Leverages CNN-based instance planar segmentation for improved plane detection.

Related

  • Project
    luigifreda/pyslam

    A hybrid Python/C++ modular framework for Visual SLAM that integrates modern AI depth prediction, semantic segmentation, and volumetric reconstruction for research and prototyping.

  • Project
    ACFR-RPG/DynOSAM

    DynoSAM is a Stereo/RGB-D Visual Odometry pipeline for Dynamic SLAM that estimates camera poses and object motions while mapping static backgrounds.

  • Project
    Tsinghua-MARS-Lab/SLAM-Former

    A Transformer-based SLAM system that unifies localization and mapping into a single model for 3D reconstruction and camera tracking from image sequences.

  • Project
    MIT-SPARK/VGGT-SLAM

    A real-time dense feed-forward scene reconstruction system for 3D mapping and localization, featuring optional open-set 3D object detection.

  • Project
    stella-cv/stella_vslam

    stella_vslam is an open‑source visual SLAM library (monocular, stereo, RGB‑D) that supports many camera models, lets you save/load maps, and offers modular APIs plus ROS integration. It’s a community fork of OpenVSLAM, actively maintained under a BSD‑2 license.