This program can run on MacOS/Ubuntu/Raspbian. Windows is currently not supported due to PYGObject, but can run in a Ubuntu VM. See Ubuntu Install for instructions. Below are the install instructions for Mac OS and Ubuntu.
https://github.com/IntelRealSense/librealsense/blob/master/doc/installation_raspbian.md
https://www.learnopencv.com/install-opencv-4-on-raspberry-pi/
https://gist.github.com/neilyoung/8216c6cf0c7b69e25a152fde1c022a5d
https://docs.mopidy.com/en/v0.8.0/installation/gstreamer/
https://www.amazon.com/gp/product/B07WQW6H9S
Install homebrew, follow the steps from here. Make sure that xcode is installed and up to date before proceeding.
Install Python 3.6.8. This is the version that is verified to work, but it may work with other versions of Python.
Install PyGObject by following the steps here. If you're using a Python version manager like pyenv, follow the instructions here. We recommend following the instructions to create a development environment with PyGObject to make it easier to work with Python 3.6.8: https://pygobject.readthedocs.io/en/latest/devguide/dev_environ.html next. Homebrew might be able to install it as well with brew install pygobject3 gtk+3.
Install dependencies for building both realsense and opencv:
brew install cmake libusb pkg-config
brew install gstreamer
brew install gst-plugins-good gst-plugins-bad gst-libav
brew install --cask apenngrace/vulkan/vulkan-sdk
xcode-select --installDownload librealsense: git clone https://github.com/IntelRealSense/librealsense.git
Follow the guide on Intel Realsense's github up until the final cmake command (Step 4). Use the command below instead to build realsense for Python 3.6.8. If you're using a virtual environment or a different version of Python, be sure to edit -DPYTHON_EXECUTABLE to match the path to the correct version of python.
sudo cmake .. \
-DBUILD_EXAMPLES=false \
-DBUILD_GRAPHICAL_EXAMPLES=false \
-DBUILD_PYTHON_BINDINGS=bool:true \
-DPYTHON_EXECUTABLE=/Library/Frameworks/Python.framework/Versions/3.6/bin/python3.6
sudo make -j2
sudo make installAfter sudo make install has finished, note the paths to the files pyrealsense2.cpython-36m-darwin.so and pybackend2.cpython-36m-darwin.so in your terminal window's output. These may look something like /Library/Python/3.8/site-packages/pyrealsense2/pyrealsense2.cpython-36m-darwin.so. You will need these paths in the next section.
git clone https://github.com/volumetricperformance/raspberry-capture-kit.git
cd raspberry-capture-kit
git fetch
git checkout releaseContinue from here with the two file paths you noted in the previous step.
cp /PATH/TO/site-packages/pyrealsense2/pyrealsense2.cpython-36m-darwin.so pyrealsense2.so
cp /PATH/TO/site-packages/pyrealsense2/pybackend2.cpython-36m-darwin.so pybackend2.so
python3.6 -m pip install virtualenv
python3.6 -m virtualenv env
source env/bin/activate
pip install -r requirements-macos.txtWith the virtual environment active in the repo directory cloned in the previous section, run the command pip install opencv-contrib-python.
Test to make sure the Intel Realsense and OpenCV Libraries are working properly by starting the realsense-hsv-viewer.py program with the camera connected. It should open a window and show a preview of the Realsense camera.
This was tested and verified on Ubuntu 20.04 LTS, but should work on many Ubuntu versions. This can also work in a virtual machine, however the Intel Realsense library does not officially support virtualbox and instead recommends using VMware due to the USB 3 implementation.
sudo apt update
sudo apt install git cmake
sudo apt-get install build-essential python3-dev libsqlite3-dev libgdbm-dev zlib1g-dev libbz2-dev sqlite3 tk-dev zip libssl-dev liblzma-dev libreadline-dev libffi-dev
sudo apt install libgirepository1.0-dev gcc libcairo2-dev pkg-config python3-dev gir1.2-gtk-3.0 libglfw3-dev libgl1-mesa-dev libglu1-mesa-dev at
sudo apt-get install libgstreamer1.0-0 gstreamer1.0-plugins-base gstreamer1.0-plugins-good gstreamer1.0-plugins-bad gstreamer1.0-plugins-ugly gstreamer1.0-libav gstreamer1.0-doc gstreamer1.0-tools gstreamer1.0-x gstreamer1.0-alsa gstreamer1.0-gl gstreamer1.0-gtk3 gstreamer1.0-qt5 gstreamer1.0-pulseaudioLoosely following the guide here, input the following commands.
wget https://www.python.org/ftp/python/3.6.8/Python-3.6.8.tgz
tar xf Python-3.6.8.tgz
cd Python-3.6.8
./configure --enable-optimizations --prefix=/opt/python3.6.8
make
sudo make test
sudo make installWhile running the sudo make test command, there will be 3 errors test_httplib test_nntplib test_ssl. Everything will still work as intended regardless.
Type in this command sudo ln /opt/python3.6.8/bin/python3.6 /usr/local/bin/python3.6 to make it easier to run python3.6 from the terminal.
Follow the Intel Realsense guide up intil the cmake step (Step 4). Instead, replace that step with:
sudo cmake .. -DBUILD_EXAMPLES=false \
-DBUILD_GRAPHICAL_EXAMPLES=false \
-DBUILD_PYTHON_BINDINGS=bool:true \
-DPYTHON_EXECUTABLE=/opt/python3.6.8/bin/python3.6
sudo make -j2
sudo make installgit clone https://github.com/volumetricperformance/raspberry-capture-kit.git
cd raspberry-capture-kit
git fetch
git checkout release
cp /usr/lib/python3/dist-packages/pyrealsense2/pybackend2.cpython-36m-x86_64-linux-gnu.so pybackend2.so
cp /usr/lib/python3/dist-packages/pyrealsense2/pyrealsense2.cpython-36m-x86_64-linux-gnu.so pyrealsense2.so
python3.6 -m pip install virtualenv
python3.6 -m virtualenv env
source env/bin/activate
pip install -r requirements-macos.txtTest to make sure the Intel Realsense and OpenCV Libraries are working properly by starting the realsense-hsv-viewer.py program with the camera connected. It should open a window and show a preview of the Realsense camera.
In order to run the main capture.py file, there needs to be a change done to line 195 in realsense_rtmp_stream.py. Change omxh264enc to x264enc.
Start the capture kit software with python capturekit.py.
The terminal will show an address similar to this: Running on http://0.0.0.0:5000
Point a web browser this address and follow the instructions in this article starting at the section "Setting up the stream."