Multi-Camera Person Tracking with Consistent IDs – Real-Time

[marix-9]

I’m trying to build a system in Python that uses more than one camera to track people. I want to make sure that if the same person appears in both cameras (where their views overlap), they get the same ID. If someone is only in one camera, they can have a different ID.

I’m using an RTX 3050 GPU, and I need the system to run smoothly with a steady frame rate, because it’s for a real-time use case where speed and accuracy really matter.

Can anyone suggest how I can do this? Any tools, libraries, or ideas would be really helpful.

[sophwats]

Hi @marix-9 I recommend you take a look at the NVIDIA Multi-Camera Tracking AI Workflow.

Best,

Sophie

[Rajana-Kavinda]

It is only available for enterprise GPUs. Isn’t it?

[sarodeanil35]

We are exploring the possibility of implementing NVIDIA’s multi-camera tracking solution and would like to understand its feasibility and performance in real-world scenarios. Could you please provide more details on:

• The accuracy and reliability of the solution in practical deployments.

• Any existing use cases or references from your clients where this solution has been successfully implemented?

[sophwats]

Hi @sarodeanil35

For accuracy and other KPIs, please take a look at this documentation page: https://docs.nvidia.com/mms/text/MDX_Multi_Camera_Tracking_KPI.html

For use cases, you can refer to https://www.nvidia.com/en-us/use-cases/ai-powered-multi-camera-tracking/.
This is still an early access product, so specific customer references will be published later, but there’s one public example here: https://blogs.nvidia.com/blog/computex-metropolis-nim/

Best,

Sophie