rainbroken/lio

How rainbroken/lio works

IteraLIO is a ROS 2 LiDAR–inertial odometry and mapping system. Its central workflow accepts sensor measurements, synchronizes and prepares scans, estimates motion by scan-to-map correction, updates local and global maps, and publishes odometry and map outputs. Optional backend keyframes and loop verification support pose-graph correction; evaluation records and run-comparison tools provide a separate analysis branch. The main pipeline and selected backend/loop internals are sampled; links for other subsystems follow the documented capabilities and repository layout, so uncertain internal call paths are omitted.

Main entry points: lio_pipeline.cpp LIO pipeline · itera_lio_node.cpp ROS node · input_adapter.cpp ROS input adapter · compare_lio_runs.py Run comparison

Components 27

ROS input/output

Livox adapter src/interfaces/src/livox_to_pointcloud2_node.cpp
ROS input adapter src/itera_lio/src/ros/input_adapter.cpp
Measurement synchronizer src/itera_lio/src/sensor/measurement_synchronizer.cpp
LiDAR parser src/itera_lio/src/sensor/lidar_parser.cpp
ROS node src/itera_lio/src/itera_lio_node.cpp
Output publisher src/itera_lio/src/ros/publishers/output_publisher.cpp

Odometry frontend

LIO pipeline src/itera_lio/src/system/lio_pipeline.cpp
LIO processor src/itera_lio/src/system/lio_processor.cpp
Iterated ESKF src/itera_lio/src/estimator/iterated_eskf.cpp
LiDAR correction src/itera_lio/src/estimator/lidar_update.cpp
Scan deskewer src/itera_lio/src/frontend/deskewer.cpp
Scan-to-map registration src/itera_lio/src/frontend/scan_to_map.cpp

Mapping

Local map store src/itera_lio/src/frontend/ikd_map.cpp
Local map manager src/itera_lio/src/frontend/local_map_manager.cpp
Global map archive src/itera_lio/src/frontend/global_map.cpp
Colored map src/itera_lio/src/mapping/color/colored_global_map.cpp
PCD writer src/itera_lio/src/output/pcd_writer.cpp

Backend and loops

Backend manager src/itera_lio/src/backend/backend_manager.cpp
Pose graph optimizer src/itera_lio/src/backend/pose_graph_optimizer.cpp
Loop closure manager src/itera_lio/src/loop/loop_closure_manager.cpp
Keyframe cloud store src/itera_lio/src/loop/keyframe_cloud_store.cpp

Evaluation

Evaluation logger src/itera_lio/src/output/evaluation_logger.cpp
Run comparison tools/compare_lio_runs.py
Run records
Comparison report

Other components

LiDAR and IMU
ROS user

Connections 29

LiDAR and IMU → Livox adapter: publishes scans
Livox adapter → ROS input adapter: adapts messages
LiDAR and IMU → ROS input adapter: sends measurements
ROS input adapter → Measurement synchronizer: buffers measurements
Measurement synchronizer → LiDAR parser: supplies scans
ROS node → ROS input adapter: receives input
ROS node → LIO processor: processes frames
LIO processor → LIO pipeline: runs odometry
LIO pipeline → Scan deskewer: prepares scan
LIO pipeline → Scan-to-map registration: registers scan
LIO pipeline → LiDAR correction: corrects state
LiDAR correction → Iterated ESKF: updates estimate
Scan-to-map registration → Local map store: queries map
LIO pipeline → Local map manager: updates window
Local map manager → Local map store: prunes points
LIO pipeline → Global map archive: archives points
LIO pipeline → Backend manager: adds keyframes
LIO pipeline → Loop closure manager: submits keyframes
Loop closure manager → Keyframe cloud store: stores clouds
Backend manager → Pose graph optimizer: optimizes graph
LIO pipeline → Colored map: optional coloring
ROS node → Output publisher: publishes results
Output publisher → ROS user: emits ROS outputs
LIO pipeline → PCD writer: exports map
ROS node → Evaluation logger: records run
Evaluation logger → Run records: writes metrics
ROS user → Run comparison: selects runs
Run records → Run comparison: reads records
Run comparison → Comparison report: writes report

Drawn by GitDiagram on September 29, 2026 from the repository's file tree, README and source files. Read as Markdown ·