Oh-My-LOAM is a ROS-free implementation of LOAM (J. Zhang and S. Singh. LOAM: Lidar Odometry and Mapping in Real-time). This implementation is modified from A-LOAM.
Comparing with A-LOAM, this implementation has following features:
- ROS-free: it can run without ROS environment
- Multi-threading instead of multi-process: more deterministic
- Higher code quality: more readable and easier to understand/modify
Install dependences (listed below).
Clone this repository
Compile:
mkdir build && cd build
cmake ..
make -j6
Although Oh-My-LOAM is ROS-free, running it with ROS bag as input is the simplest way. We'll take nsh_indoor_outdoor.bag as example. You can download this bag from google drive or baidupan (提取码:9nf7).
Launch Oh-My-LOAM:
./devel/lib/oh_my_loam/main_rosbag ../configs/config_nsh_indoor_outdoor.yaml
Play ROS bag (in a new terminal):
ros play nsh_indoor_outdoor.bag
Launch Oh-My-LOAM:
./devel/lib/oh_my_loam/main_noros ../configs/config_nsh_indoor_outdoor.yaml xxxxxx
Please replace xxxxxx
with the directory that contains the input point cloud files with tree structure like following:
xxxxxx
├── frame00000.pcd
├── frame00001.pcd
├── frame00002.pcd
├── frame00003.pcd
├── frame00004.pcd
├── ...
Currently only .pcd
format is supported.
You can modify examples/main_noros.cc
to add support for other point cloud formats.
Tested on ubuntu 16.04/18.04/20.04.
If cannot find std::filesystem error is encountered during your compiling, please upgrade your compiler.
We recommend g++-9
(or higher version).
Only examples/main_rosbag.cc
needs ROS. You can exclude it from compiling by modifying examples/CMakeLists.txt
.
sudo apt install libeigen3-dev
sudo apt install libpcl-dev
Follow g3log to install.
sudo apt install libyaml-cpp-dev
sudo apt install libceres-dev