rsasaki0109/lidar_slam_ros2

What it solves

Most LiDAR SLAM (Simultaneous Localization and Mapping) systems only provide a trajectory and a point cloud. This project bridges the gap between raw SLAM output and usable autonomous driving maps by generating Autoware-compatible map bundles, including point cloud maps, projector information, and automatically generated lanelet2 drivable lanes.

How it works

The system uses a two-stage pipeline: a frontend called RKO-LIO for LiDAR-inertial odometry and a backend called graph_based_slam for loop closure and graph optimization. It can process rosbag2 files to create maps and includes specialized presets for challenging environments like tunnels or fog. Additionally, it features a camera-coloring pipeline that projects camera pixels onto the LiDAR point cloud by refining camera poses against the SLAM geometry.

Who it’s for

It is designed for robotics developers and autonomous vehicle engineers who need to turn raw sensor data (from Ouster, Velodyne, RoboSense, or Livox sensors) into high-definition maps that can be directly loaded into Autoware.

Highlights

• Autoware Integration: Directly outputs pointcloud_map/ and map_projector_info.yaml for seamless loading.
• Lanelet2 Generation: Automatically creates drivable lanelets from the SLAM trajectory.
• Degeneracy Resilience: Includes specific presets for mapping self-similar tunnels and fog-heavy environments.
• Camera Coloring: Supports projecting RGB data onto 3D maps with spatiotemporal calibration refinement.
• Deterministic Mapping: Ensures byte-identical trajectories and submaps under release gates for consistent results.