We want to record video by ffmpeg and do inference (such as YOLO) at same time.
But we can’t open /dev/video by two application at same time.
Could you provide any suggestion ?
Hi,
We would suggest use gstreamer, so that you can use tee and queue plugins to duplicate the source into different sinks. And can use DeepStream SDK to run deep learning inference.
The example of using tee and queue:
Gstreamer TCPserversink 2-3 seconds latency - #4 by Bazziil
In the server command, the source is duplicated into video preview and UDP streaming.
For using DeepStream SDK, please check
This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.