UZ-SLAMLab/ORB_SLAM3

Health v2 · maintenance only

Adoption not part of the score

8986 stars · 3163 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

ORB-SLAM3 is a real-time SLAM library supporting Visual, Visual-Inertial, and Multi-Map SLAM with monocular, stereo, and RGB-D cameras using pin-hole and fisheye lens models. It is a C++ research library from the University of Zaragoza, widely used as a state-of-the-art baseline in visual SLAM research.

Use cases

• run visual SLAM on monocular, stereo, or RGB-D camera feeds
• perform visual-inertial SLAM fusing camera and IMU data
• build sparse 3D maps for robot localization and navigation
• evaluate SLAM accuracy on EuRoC and TUM-VI datasets
• support fisheye camera models in SLAM pipelines
• implement multi-map (Atlas) SLAM with map merging

When to choose

• you need a proven, accurate real-time visual or visual-inertial SLAM system
• you are doing robotics or AR research and want a strong baseline
• you work with standard or fisheye cameras and optional IMU
• you need multi-map robustness for tracking loss and relocalization

When to avoid

• you need dense or semantic mapping rather than sparse feature-based SLAM
• you want a maintained product with frequent updates and commercial support (GPL-3.0, research code)
• you need a plug-and-play solution without C++ build and OpenCV/Pangolin dependencies
• your project requires a permissive license incompatible with GPL-3.0

Facets

library · maturity maintenance

computer-vision simulation machine-learning robotics computer-vision autonomous-vehicles cpp cross-platform slam visual-slam visual-inertial orb-features robotics camera-tracking mapping linux

Member repositories

For agents

markdown · JSON · MCP: product_card(name="UZ-SLAMLab/ORB_SLAM3")