hku-mars/LiDAR_IMU_Init

Health v2 · maintenance only

Adoption not part of the score

1491 stars · 253 forks observed · 2026-08-28

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

LI-Init is a robust, real-time initialization method for LiDAR-inertial systems that calibrates temporal offset, extrinsic parameters, gravity vector, and IMU bias without any targets or extra sensors. It includes a robust LiDAR odometry (FAST-LO) modified from FAST-LIO2 and supports multiple mechanical spinning and solid-state LiDAR types.

Use cases

• calibrate extrinsic parameters between a LiDAR and IMU without hardware setup
• estimate time offset between LiDAR and IMU sensors in real time
• initialize a LiDAR-inertial SLAM system robustly
• run robust LiDAR odometry with Hesai, Velodyne, Ouster, or Livox sensors
• integrate initialization into FAST-LIO2 pipelines
• estimate gravity vector and IMU bias for sensor fusion

When to choose

• you need targetless LiDAR-IMU extrinsic and time offset calibration
• you are building a SLAM or odometry system with LiDAR and IMU inputs
• you use Livox solid-state or common spinning LiDARs on Ubuntu with ROS
• you want a robust initialization module for FAST-LIO2

When to avoid

• you need camera-LiDAR or other multi-sensor calibration
• you work outside ROS/Ubuntu environments
• you need a fully integrated SLAM solution rather than an initialization module
• your project requires a permissive license since this is GPL-2.0

Facets

library · maturity active

robotics simulation math parser robotics autonomous-vehicles developer-tools cpp lidar imu calibration slam ros sensor-fusion initialization fast-lio odometry time-offset-calibration algorithms linux

Member repositories

For agents

markdown · JSON · MCP: product_card(name="hku-mars/LiDAR_IMU_Init")