Skip to main content

NVIDIA Jetson Platform Integration

Migrate OpenEB to Nvidia Jetson Platform​

This guide demonstrates how to migrate OpenEB to Nvidia Jetson platform using Nvidia Orin Jetpack 5.0.2 (L4T 35.1) as an example.

Reference Documents​

1. Preparation Document​

The patch file of OpenEB plugin: sensing_imx646.diff

Download it from: https://autosensee.feishu.cn/drive/folder/LS9hffxYvl6fAXdF2wmcyIZCn2g

2. Install Dependent Packages​

System Dependencies​

$ sudo apt update
$ sudo apt -y install apt-utils build-essential software-properties-common wget unzip curl git cmake
$ sudo apt -y install libopencv-dev libboost-all-dev libusb-1.0-0-dev
$ sudo apt -y install libhdf5-dev hdf5-tools libglew-dev libgifw3-dev libgifww3-dev libcanberra-gtk-module ffmpeg
$ sudo apt-get install libprotobuf-dev protobuf-compiler
$ sudo apt -y install libgtest-dev libgmock-dev

Python Libraries​

$ sudo apt -y install python3-pip python3-distutils
$ sudo apt -y install python3.X-dev  # where X is 8, 9 or 10 depending on your Python version (3.8, 3.9 or 3.10)
$ python3 -m pip install pip --upgrade
$ python3 -m pip install "opencv-python==4.5.5.64" "sk-video==1.1.10" "fire==0.4.0" "numpy==1.23.4" pandas scipy h5py
$ python3 -m pip install jupyter jupyterlab matplotlib "ipywidgets==7.6.5" pytest command_runner

Install pybind11 (version >= 2.6.0)​

$ wget https://github.com/pybind/pybind11/archive/v2.6.0.zip
$ unzip v2.6.0.zip
$ cd pybind11-2.6.0
$ mkdir build && cd build
$ cmake .. -DPYBIND11_TEST=OFF
$ cmake --build .
$ sudo cmake --build . --target install

Install PyTorch 1.13.1​

$ python3 -m pip install "numba==0.56.3" "profilehooks==1.12.0"
$ python3 -m pip install "pytorch_lightning==1.8.6" "tqdm==4.63.0" "kornia==0.6.8"

3. Compile and Install OpenEB​

Download the OpenEB source code. The current version is v4.3.0:

$ git clone https://github.com/prophesee-ai/openeb.git --branch 4.3.0

Copy the patch file sensing_imx646.diff to hal_psee_plugins directory and apply the patch:

$ git apply sensing_imx646.diff

Compile and install:

$ cd openeb
$ mkdir build && cd build
$ cmake .. -DCMAKE_BUILD_TYPE=Release -DBUILD_TESTING=OFF
$ cmake --build . --config Release -- -j 'nproc'
$ sudo cmake --build . --target install

Add the following environment variables to the end of your ~/.bashrc file:

export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/lib
export HDF5_PLUGIN_PATH=$HDF5_PLUGIN_PATH:/usr/local/hdf5/lib/plugin

Make environment variables effective:

$ source ~/.bashrc

4. Connect and Test Equipment​

4.1 Connect Equipment​

Connect EVS camera to the USB port of NVidia Orin device via USB cable.

View the device through the lsusb command:

$ lsusb

You should see output containing "SENSING USB3 EVS CAMERA" in the list.

4.2 View Device Information​

$ metavision_platform_info

This will display detailed information about the connected camera, including:

  • Available data encoding formats
  • Current data encoding format
  • Connection type
  • Camera build date and version
  • Sensor name and serial number
  • Available device configuration options

4.3 Run Sample Viewer​

$ metavision_viewer

This will launch the Metavision viewer application where you can visualize the event-based camera output.

What's next?​