Clone the Git repository and then use pip:
git clone git@github.com:psiexperiment/psivideo
pip install -e ./psivideo
This should automatically pull all dependencies you need (including OpenCV, PyAV, and websockets).
This comes with both the server and client. You'll primarily be working with the client. To test the client, run either demo_sync_client.py or demo_async_client.py in the scripts folder.
To quit, select the video window and hit thq "q" key.
Five separate threads are used for running the server:
- Video capture. This is very simple and just grabs the next image from the camera and puts it into thread-safe queue that is read by the video process thread.
- Video process. This currently just moves the image from the video processing queue to the video writing queue (if a client has requested the video be saved) as well as the online image display thread.
- Video display. This displays a small window on the screen. This is not necessarily updated at the same rate as the video capture (it updates as fast as possible, but there may be dropped frames).
- Video write. This writes the video to an AVI file.
- Video communication. This listens for incoming connections from a websocket client on port 33331. Payload is json.
Want to customize the program?
To add new processing features, overide Video.process to do what you want. Be sure to check the camera frame rate before and after implementing your changes. If the frame rate drops once the changes are made, you need to set up the capture process to run in a separate process (rather than thread).