How to run multi thread in YOLOv3 on nano?

Autonomous Machines Jetson & Embedded Systems Jetson Nano
1

Hello all
Now, i am using YOLOv3 to do the object detection on Jetson nano. But i want to change the code to achieve the goal: If detect the person then control one relay to cut the power 2 seconds. But i don’t know how to change the code. I try to create threat using pthread_create() and sleep 2 seconds. But it have influence on real-time detection. Is there somebody could help me?Looking forward to your reply!
Thank you in advance!!!

2

Hi,

Maybe you can separate the inference part with a thread.

Generate a thread when a frame arrives and terminate it after inferencing.
Once a person detected, freeze the process with two seconds.

Thanks.

3

Hello AastaLLL
I am not fimiliar with how to separate the inference. Do you have another way to achieve my goal?

4

Hi,

What kind of source do you use?
Is it darknet? Or TensorRT? Or others?

Thanks.

5

hello
I use darknet.

6

Hi,

It looks like darknet executes the inference with the function call like this:
https://github.com/pjreddie/darknet/blob/master/examples/detector.py#L19

r = dn.detect(net, meta, "data/eagle.jpg")

So you can add some mechanism on this to achieve temporally sleep.
Thanks.