uzh-rpg/rpg_svo_pro_open

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

1601 stars · 416 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

SVO Pro is a C++ implementation of Semi-direct Visual Odometry from the Robotics and Perception Group at University of Zurich, supporting monocular and stereo setups with perspective and fisheye cameras. It extends to visual-inertial odometry and full visual-inertial SLAM with sliding-window optimization, global bundle adjustment via iSAM2, and loop closure using DBoW2.

Use cases

• estimate camera trajectory from monocular or stereo video
• run visual-inertial odometry on drone or robot sensor data
• build a globally consistent SLAM map with loop closure
• benchmark SLAM pipelines on the EuRoC dataset
• handle fisheye or catadioptric camera models in odometry
• control camera exposure actively for robust tracking

When to choose

• you need a fast, research-grade visual or visual-inertial odometry front-end in C++
• you want full SLAM with global bundle adjustment and loop closure
• you work with ROS and standard datasets like EuRoC
• you need support for fisheye and catadioptric camera models

When to avoid

• you need a permissively licensed solution for commercial products (GPLv3)
• you want a plug-and-play consumer SLAM app rather than a research codebase
• you need RGB-D or LiDAR-based SLAM
• you require long-term maintenance and frequent updates

Facets

library · maturity maintenance

computer-vision simulation graphics robotics autonomous-vehicles computer-vision cpp visual-odometry slam visual-inertial-odometry loop-closure bundle-adjustment robotics-research euroc algorithms linux ros

Member repositories

For agents

markdown · JSON · MCP: product_card(name="uzh-rpg/rpg_svo_pro_open")