Simultaneously viewing an inferenced camera feed while saving inferenced video

How might one go about inferencing a camera feed and saving a video result of said inferencing to a video?

I’ve seen a few solutions for viewing with overlay in a display and then having the raw video saved but I’m wondering is there a way to view the overlay in a display and then have a saved video with the overlay included.

We suggest use deepstream-app. You can enable multiple sinks in config file. For your use-case, you can enable [sink0] for overlay and [sink1] to save to a video file. Please check
NVIDIA Metropolis Documentation

would there be any way to do this using jetson inference and utils? in other words, in jetson utils can I have more than one video output?
for instance if I did:

display = jetson.utils.videoOutput("display://0")
outFile = jetson.utils.videoOutput("video.mp4")

Hi @takice2328, jetson.utils.videoOutput will automatically make a display://0 output in addition to the primary output. So if you encode it to video file, you should also see it display on screen.

This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.