Retrieve annotated video for re-encoding and re-transmission


I’m having this pipeline successfully running on a Jetson Nano

nvstreammux -> nvinfer -> nvtracker -> nvmultistreamtiler -> nvvideoconvert -> nvdsosd -> nvegltransform -> nveglglessink

I’m colourizing the detected object borders in a src probe handler attached to nvtracker. I would like to grep the annotated raw video and “tee” it to either a GStreamer appsink element for manual processing or directly to a H.264 hardware encoder (maybe nvv4l2h264enc) for provisioning to AIORTC and WebRTC transmission to the outside.

Anybody having experiences with this kind of handling, able to share?

TIA and Happy Easter for those who celebrate that

Found i:

Basically the same, but different :)

nvstreammux -> nvinfer -> nvtracker -> nvmultistreamtiler -> nvvideoconvert -> nvdsosd -> nvvideoconvert -> capsfilter -> nvv4l2h264enc -> filesink

The nvvideoconvert after nvdsosd links to a capsfilter video/x-raw(memory:NVMM), format=I420, which prepares the input for the encoder. The filesink is just a placeholder for whatever comes in terms of WebRTC transmission. filesink saves raw H.264 packets and is an ideal input for AIORTC.

There is no update from you for a period, assuming this is not an issue anymore. Hence we are closing this topic. If need further support, please open a new one. Thanks

Yes. You can use this hardware encoder plugin.

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.