Video object detection on Jetson Nano for beginners

Hi, I am new to Jetson nano, and I will like to get some pointers on whether the use cas I have in mind is possible with a jetson nano.

I want a system that can receive a video feed from a full HD IP camera and it will simply sound a buzzer every time a certain object (trained from an image library) appears on the video feed.

being the camera full hd, how many images per second (just a rough idea) do you recon the basic jetson nano will be able to process?

is this project feasible? can you point me to some software that i could use for this application?


Please try DeepStream SDK. You can install the package through SDKManager. And the package is installed to


The documentation:
NVIDIA Metropolis Documentation

If you have queries, please create topic in
DeepStream SDK - NVIDIA Developer Forums