Here you will find a suite of tools designed to work with the ColoRadar Dataset.
Login to the Github container Registry:
docker login ghcr.io
Pull an image:
docker pull ghcr.io/arpg/coloradar-lib:{<X.Y>}{-<ros_distro>}
Example:
docker pull ghcr.io/arpg/coloradar-lib:12.6-jazzy
ros_version | OS | Python | Cuda |
---|---|---|---|
humble | Ubuntu 22.04 Jammy | 3.10 | 12.6, 12.8, none |
jazzy | Ubuntu 24.04 Noble | 3.12 | 12.6, 12.8, none |
none | Ubuntu 24.04 Noble | 3.12 | 12.6, 12.8, none |
- OS: Linux/Mac
- Docker: Version 20.10+
- NVIDIA GPU Drivers (Optional): Compatible with the chosen CUDA version
- NVIDIA Container Toolkit (Optional): For GPU support in Docker
Run the provided script to build an image for your desired ROS and CUDA versions
python3 build.py --ros <ros_distro> --cuda <cuda_version>
<ros_distro>
: ROS distribution (e.g.,noetic
,humble
,jazzy
). Leave empty or usenone
for no ROS.<cuda_version>
: CUDA version inX.Y
format (e.g.,12.4
). Usenone
to skip CUDA.
Examples:
-
Build with ROS Noetic with CUDA 12.4:
python3 build.py --ros noetic --cuda 12.4
-
Build ROS Humble while detecting the local CUDA version:
python3 build.py --ros humble
-
Build without ROS or CUDA:
python3 build.py --ros none --cuda none
Note: the CUDA-ROS base images are pulled from my dockerhub.
mkdir build && cd build
cmake ..
make
- Linux/Mac
- GCC v.10-12
- BOOST 1.78+
- open-mpi
- VTK
- openCV
- PCL v.1.12+
- octomap
- Yaml-cpp
- (optional) PyBind11 v.2.6+
./build/coloradar_tests
python3 tests/test_bindings.py
@article{Harlow2024ARXIV,
title = {ColoRadar+: An extension of the dense
millimeter-wave radar dataset ColoRadar},
author = {Kyle Harlow, Doncey Albin, Miles Mena},
journal = {arXiv preprint arXiv:####.#####},
year = {2024},
}