python-docker-opencv-demo
Instead of using webcam capture and GUI output, this sample reads from a file and outputs to a new file.
To build the container:
$ git clone https://github.com/brianshaler/python-docker-opencv-demo.git
$ cd python-docker-opencv-demo
$ docker build -t pdocv .
Now you can docker run [opts] pdocv [cmd]. For example:
$ docker run --rm -v $(pwd)/data:/data pdocv python ./main.py ./video/input.webm /data/output.webm
The above command creates/links local directory ./data to
/data inside the container (-v $(pwd)/output:/data), and
processes the project's sample video (video/input.webm) and
outputs it to ./data/output.webm in the project directory.
If you want to modify main.py, either make changes and run
docker build -t pdocv . again, OR link your working copy
with -v $(pwd):/usr/src/app (before or after the other -v
used above, but MUST be after docker run and before
pdocv). Then, your current copy of main.py would be used
instead of the one in the image.
e.g. Make a change to main.py in your project directory,
run:
$ docker run --rm -v $(pwd):/usr/src/app pdocv python ./main.py ./video/input.webm ./data/output2.webm
(note: ./data is a relative path in this example, since
/data was not mounted)
You should then see your changes reflected on in the video
located at ./data/output2.webm.