Jetson / PC demo

Docker configuration

Docker should be configured to be able to access the GPU throught nvidia-containers. To install nvidia-container-runtime:

sudo apt-get install nvidia-container-runtime

If docker is not yet configured to use nvidia-container-runtime, it can be done with:

sudo tee /etc/docker/daemon.json <<EOF
{
    "runtimes": {
        "nvidia": {
            "path": "/usr/bin/nvidia-container-runtime",
            "runtimeArgs": []
        }
    }
}
EOF

Then restart docker service:

 sudo systemctl daemon-reload
 sudo systemctl restart docker

Building and installation instructions

0) Clone the private git repo

git clone https://[email protected]/oriel/doodlenet_heliaus_demo.git

1) Build docker image

Go to the path of the cloned repository:

cd doodlenet_heliaus_demo

a) Build for PC

docker build -f Dockerfile_pc_ros2 -t "doodlenet_ros2:beta" .

b) Build for jetson (will only work if building from a jetson - L4T 32.6.1 [ JetPack 4.6 ])

docker build -f Dockerfile_jetson_ros2  -t "doodlenet_ros2:beta" .

2) Run docker image:

Run this bash script to allow gpu usage, host network, X graphic server:

bash runDocker.sh

3) Running ros2 demo inside docker image

python3 autosens_demo.py

3.1) Running ros2 demo with custom ROS2 topics for input cameras

Note that the demo with default parameters use "source_images_rgb" and "source_images_lwir" as input ROS2 topics. You may use custom input topics for the Autosens demo with:

python3 autosens_demo.py --rgb_topic /camera/flea/left/aligned --lwir_topic /camera/smartIR640/cam_0008/ir_frame_enhanced/recified

You may check other available customisable parameters with:

python3 autosens_demo.py --help