Python Deepstream 5.0 app with TX2 onboard and raspberry pi camera

• Hardware Platform (Jetson / GPU): Nano and TX2
• DeepStream Version: 5.0 GA
• JetPack Version: 4.4
• TensorRT Version: 7.1

How can I change the input source in deepstream-test1 from video file to live onbord camera in TX2 or raspberry pi camera in Nano.

Thank you.

You can use deepstream-app sample, and samples/configs/deepstream-app/source1_csi_dec_infer_resnet_int8.txt is for CSI camera.

Thank you for the answer.
I understand that we can use config file and start the application via
deepstream-app -c source1_csi_dec_infer_resnet_int8.txt

However, my task is required to work python because I need to extract the bounding box and other metadata to do some post processing.

In the example the file source gst pipeline start with this

print("Creating Source \n ")
source = Gst.ElementFactory.make("filesrc", "file-source")
if not source:
    sys.stderr.write(" Unable to create Source \n")

So, is it possible to start the gst pipeline with the CSI camera in python?

Edit: I have found the solution!


Have you already done the changes and works ok?
Could you share what changes are needed?
Thanks @DarkEvilz