Offloading Data for Deep learning inference on Jetson Nano

Autonomous Machines Jetson & Embedded Systems Jetson Nano
1

Hi, is there anyway to offload the data captured by camera connected to light weight Embedded devices let say raspberry or such kind and offload to jetson present at distance.

Main application is drone, so cameras (integrated to embedded device) on drones would capture data and send it to the jetson nano, then jetson nano would do the required processing (DL inference) and send the message back to the drone.

Note that data will be offloaded and data can be of any kind such as video or image for DL inference of classification and so on…

2

Any updates team?

4

Hi,

Could you share more about your usecase?

Your drone isn’t built on the Nano, is that correct?
Drone and Nano are two different devices but do you want to send the data captured by Drone to Nano to use GPU for inferencing?

If yes, you can try RTSP with Deepstream SDK:

Thanks.

5

Hi thanks for the reply.

Yes @AastaLLL you are right. I want to stream my drone’s camera video onto the ground PC connected to jetson for Computer vision applications.

I would want to build gstreamer pipeline to handle all the stuff, and also the drone will have its flight controller so that i can control the drone manually but the camera should simple do live streaming for Computer vision applications.

6

Any update team?

Also check this below link for reference. I want somthing like this which is in real time

7

Hi,

You can try Deepstream SDK.
Deepstream can be used for object detection with the RTSP input stream.

Thanks.

9

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