Thanks to hku mars lab chunran zheng for the open source excellent work

• 🔓 2025-01-23: Code released!
• 🎉 2024-10-01: Accepted by T-RO '24!
• 🚀 2024-07-02: Conditionally accepted.

If you have any questions, please feel free to contact: Chunran Zheng zhengcr@connect.hku.hk.

FAST-LIVO2 is an efficient and accurate LiDAR-inertial-visual fusion localization and mapping system, demonstrating significant potential for real-time 3D reconstruction and onboard robotic localization in severely degraded environments.

Our accompanying video is now available on Bilibili and YouTube.

FAST-LIVO2: Fast, Direct LiDAR-Inertial-Visual Odometry

FAST-LIVO: Fast and Tightly-coupled Sparse-Direct LiDAR-Inertial-Visual Odometry

We open-source our handheld device, including CAD files, synchronization scheme, STM32 source code, wiring instructions, and sensor ROS driver. Access these resources at this repository: LIV_handhold.

Our associate dataset FAST-LIVO2-Dataset used for evaluation is also available online. Please note that the dataset is being uploaded gradually.

MARS-LVIG dataset：A multi-sensor aerial robots SLAM dataset for LiDAR-visual-inertial-GNSS fusion

Ubuntu 22.04. ROS Installation.

PCL>=1.6, Follow PCL Installation.

Eigen>=3.3.4, Follow Eigen Installation.

OpenCV>=3.2, Follow Opencv Installation.

sudo apt install ros-$ROS_DISTRO-sophus

Sophus Installation for the non-templated/double-only version.

mkdir -p fast_ws/src && cd fast_ws/src git clone https://github.com/strasdat/Sophus.git cd Sophus git checkout 1.22.10 mkdir build && cd build cmake .. -DCMAKE_BUILD_TYPE=Release make -j$(nproc) sudo make install

Vikit contains camera models, some math and interpolation functions that we need.

# Different from the one used in fast-livo1 cd fast_ws/src git clone https://github.com/integralrobotics/rpg_vikit.git

Thanks to the following repositories for the code reference:

• uzh-rpg/rpg_vikit
• xuankuzcr/rpg_vikit
• uavfly/vikit

Follow livox_ros_driver2 Installation.

why not use livox_ros_driver? Because it is not compatible with ROS2 directly. actually i am not think there s any difference between livox ros driver and livox ros driver2 's CustomMsg, the latter 's ros2 version is sufficient.

Clone the repository and colcon build:

cd ~/fast_ws/src git clone https://github.com/Ryusei-Baba/FAST-LIVO2.git cd ../ colcon build --symlink-install --continue-on-error source ~/fast_ws/install/setup.bash 

Download our collected rosbag files via OneDrive (FAST-LIVO2-Dataset).

convert ROS1 rosbag to ROS2 rosbag

pip install rosbags rosbags-convert --src Retail_Street.bag --dst Retail_Street

• gitlab rosbags
• pypi rosbags

Such as dataset Retail_Street.db3, because we use livox_ros2_driver2's CustomMsg, we need to change the msg type in the rosbag file.

1. use rosbags-convert to convert rosbag from ROS1 to ROS2.
2. change the msg type of msg type in metadata.yaml as follows:

metadata.yaml

rosbag2_bagfile_information: compression_format: '' compression_mode: '' custom_data: {} duration: nanoseconds: 135470252209 files: - duration: nanoseconds: 135470252209 message_count: 30157 path: Retail_Street.db3 .............. topic_metadata: name: /livox/lidar offered_qos_profiles: '' serialization_format: cdr - type: livox_ros_driver/msg/CustomMsg + type: livox_ros_driver2/msg/CustomMsg type_description_hash: RIHS01_94041b4794f52c1d81def2989107fc898a62dacb7a39d5dbe80d4b55e538bf6d ............... .....

Do not forget to source your ROS2 workspace before running the following command.

ros2 launch fast_livo mapping_aviz.launch.py use_rviz:=True ros2 bag play -p Retail_Street # space bar controls play/pause

The source code of this package is released under the GPLv2 license. For commercial use, please contact me at zhengcr@connect.hku.hk and Prof. Fu Zhang at fuzhang@hku.hk to discuss an alternative license.