Pleorasrc in Deepstream

Please provide complete information as applicable to your setup.

• Hardware Platform (Jetson / GPU) AGX Xavier
• DeepStream Version Deepstream 5.1
• JetPack Version (valid for Jetson only) JP 4.5
• TensorRT Version Version

I am trying to modify deepstream-test4 to include Pleorasrc as the input type, so I can have the ability to stream metadata of the detections after performing inference on input from a camera. However, I am unable to get the right plugins and the format in the pipeline. Can any of you guide me with this?

Thank you!

Try to replace filesrc and decoder with the pleorasrc. Please refer below.

Hi Kesong, thank you for your reply. I am trying to implement a similar pipeline.
I have been able to get Kafka output from deepstream-test4. I am currently working on getting an RTSP output along with metadata over Kafka.

pleorasrc → nvstreammux → nvinfer → nvvidconv →
nvosd → tee → queue1 → msgconv → msgbroker
|-> queue2 → nvvidconv, x264enc, rtph264pay, udpsink

When I run this pipeline, the execution freezes at frame 4 and does not proceed. I am unable to see any RTSP output nor do I see messages over Kafka.

I set the gst_debug to 6 and I see that queue2 is getting empty for some reason. Can anyone help me to find what may be incorrect here?

Try to add more queue in the pipeline.