The question of camera multiplexing

Because of some requires, I want the camera multiplexing, and I can see the camera through the computer(using RTSP). Streams with different resolutions use different ports(like for the same camera on TX2: 1280720- RTSP port:8553 19201080-
RTSP port:8554 …).

My needs look like the sample-Frontend of L4T Multimedia API:
But this sample used L4T Multimedia API, which is a collection of lower-level APIs.
And the gstreamer-rtsp-test-launch is a high-level APIs.
I think it is feasible, but i don’t know how to do.

That’s my problem.Does anyone have any good ideas or solutions?

If you need to run deep learning inferences on the camera sources, we suggest you try Deepstream SDK.

Thanks, But I don’t run something about deep learning at present.
I want to develop an industrial camera product, the problem I’m facing right now is that switching its resolution according to different situations, Such as bandwidth. When the network quality is very good, I converte the resolution to 19201080, conversely, I converte the resolution to 1280720( even lower).
So, I wonder if there’s any good advice.

Here is a sample of combining gstreamer + tegra_multimedia_api:
For your reference.

When I run that, I get error:

nvbuf_utils: Could not get EGL display connection
Using launch string: rtspsrc location=rtsp:// ! rtph264depay ! h264parse ! nvv4l2decoder ! appsink name=mysink sync=false 
[ERROR] (NvEglRenderer.cpp:98) <renderer0> Error in opening display
[ERROR] (NvEglRenderer.cpp:154) <renderer0> Got ERROR closing display
Segmentation fault (core dumped)

If you have executed:

$ export DISPLAY=:1(or 0)

But still hit the issue. We would like to suggest you re-flash the whole package through sdkmanager.

A similar topic:

It is okay, I changed ‘export DISPLAY=:1’ to ‘export DISPLAY=:0’.