VINS-Mono / VINS-Fusion

An open-source visual-inertial localization system from HKUST that fuses a camera and an IMU to estimate pose in real time.

VINS-Mono is a monocular visual-inertial state estimator open-sourced by Shaojie Shen’s team (the Aerial Robotics Group) at the Hong Kong University of Science and Technology, published in IEEE T-RO. Using just one camera plus one IMU (inertial measurement unit, which measures acceleration and angular velocity), it computes a device’s position and orientation in real time. It works by tightly coupling pre-integrated IMU data with image feature points inside a nonlinear optimization, and includes automatic initialization, online calibration of the camera-IMU extrinsics, loop closure detection, and 4-degree-of-freedom pose graph optimization. VINS-Fusion, released in 2019, is an extended version supporting monocular+IMU, stereo, and stereo+IMU configurations, and also demonstrates fusion with GPS. Both are built on ROS and open-sourced under GPLv3, and are commonly used as a baseline localization solution for drones and mobile robots.

ExampleA quadruped robot dog is fitted with a stereo camera that has a built-in IMU; running VINS-Fusion lets it output its body trajectory in real time indoors, with no GPS available.

Also called

VINS, Visual-Inertial Navigation System

Related

Visual-Inertial Odometry · Visual SLAM · Inertial Measurement Unit · IMU Preintegration · Loop Closure Detection · Camera-IMU Calibration

Sources

VINS-Mono: A Robust and Versatile Monocular Visual-Inertial State Estimator (arXiv)
HKUST-Aerial-Robotics/VINS-Mono (GitHub)
HKUST-Aerial-Robotics/VINS-Fusion (GitHub)

As of

2019-01