Fast Object Ditect Implementation

Robotics & Edge Computing Jetson & Embedded Systems Jetson Nano
1

Hi.
I am currently trying to do Object Detection from a USB camera video.
I’m despairing because the Object Detection that I finally created inexperiencedly only works at 3 FPS.
Does anyone have any wisdom on how to make the program run more fast?

The following is a description of the current situation.
Product used:Jetson Nano B01
Jetpack ver:4.6
Python ver:3.6
Environment:OpenCV4.5 ,tensorflow2.5+nv21.6 ,ssd_keras
Model used:weights_SSD300.hdf5
Resolution:300x300

Also, if there are any mistakes in the way you made it in the first place, please point them out to me.
I will be grateful for any help you can provide.

3

Hi,

Have you tried our Deepstream SDK?
The pipeline has optimized implementation that can give you a better performance.

Thanks.

4

Dear AastaLLL

Thank you very much for your valuable input.
I have not tried the Deepstream SDK yet.
According to the catalog specs, it can detect the human body at FPS10 or higher, so I will try this one.

Thanks,

6

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