Stereovision OpenCV & DeepStream

Hello! I’m trying to make and run stereo vision app based on OpenCV library on Jetson Nano and Xavier NX. Is it possible to use it inside Deepstream docker containers or just with DeepStream SDK for higher performance?

I’m using Isaac SDK with deepstream and a GST pipeline and it’s working fine with almost any UVC camera, unfortunatly I don’t know about performance but if you find out please keep this thread updated, I’m very interested !
What stereo camera are you using ? I’ve one but the image quality is poor and I’m looking to switch to a global shutter one.
Thanks !

HI @iklemyshev ,
Sorry for long delay! does stereo vision mean there are two cameras and the two camera can generate depth informantion ?

If you can use Xavier, there is VPI SDK (VPI - Vision Programming Interface: Stereo Disparity Estimator) can calculate the depth.

But, I don’t think it’s easy to support stereo vision , do you have the planned pipeline?

Regarding to the performance, I think, the perf in docker and outside of docker should be very close.