livox-msg-pointcloud

This is a ROS package (tested in ROS noetic) to convert Livox custom message to pcl pointcloud. The conversion include two types:

1. Real-time conversion with a C++ ROS wrapper
2. A Python script (rosbag_msg_converter.py) to convert the topics inside a rosbag by postprocessing

livox_msg_pointcloud/
├── CMakeLists.txt
├── README.md
├── include
│   └── utils.h
├── launch
│   └── livox_converter.launch
├── package.xml
├── scripts
│   ├── rosbag_check.py
│   └── rosbag_msg_converter.py
└── src
    └── livox_msg_pointcloud_node.cpp

The main conversion code is inside utils.h file and ivox_msg_pointcloud_node.cpp is a ROS node wrapper. You can also directly use the header file in our coude if you don't use ROS.

Real-time Conversion

You can change the input topic name and output topic name in the launch file and then run

roslaunch livox_msg_pointcloud livox_converter.launch

or just give them as parameters in the terminal

roslaunch livox_msg_pointcloud livox_converter.launch input_topic:=/custom/input_topic  output_topic:=/custom/output_topic  scan_merge_count:=1

Parameters explained:

input_topic # the original Livox LiDAR point cloud in custom msg
output_topoic # the converted LiDAR point cloud in PCL pointcloud format
scan_merge_count #the original point cloud is in high frequency but very sparse. We can choose to low the frequency to get dense point cloud.

Conversion by Post-Processing

The script needs various libaries:

pip install open3d numpy tqdm

And here is how to run the script.

python convert_rosbag.py input.bag output.bag 2 /topic1 /topic2

TODO

• The python script is too slow now