The end customer is using a solution based on WebRTC.

The goal is for the rider to receive the coach’s words and advice in real-time.

The system currently runs on Windows. It is equipped with an IP camera that connects to a computer. This Windows computer simulates a webcam from this IP camera. Then a script initiates the session on the rider’s side by connecting from Chrome to a given address provided by the system – https://…

The video conferencing system, which should not be touched at the moment, then starts.

The idea is as follows:

Replace the Linux PC with a Windows system.
First step: connect the IP camera and virtualize a fake USB camera.
Second step: verify that the system still works.
Third step (the goal): implement tracking and possible zoom (we are trying to improve). The problem is that I don’t see how to achieve this between the virtual USB camera and the browser. The idea is to manipulate the RTSP stream. I am more familiar with Jetson systems, where setting up tracking (SORT) with zoom is not a problem (thanks to GPUs). If we can set up tracking first, we will be on the right track, and then we can consider zoom later.
For my part, this can be done quite quickly. We have access to the Raspberry Pi via ngrok and an IP camera.

I propose that you work on a machine at your place, push the code to a GitHub repository that I will create, and then we can transfer everything to the client’s machine.

You must be able to start to work on it asap.

More information :
tracking sort will be perfect
https://github.com/abewley/sort

rtsp from ip camera – to fake a usb camera .

https://stackoverflow.com/questions/26590174/convert-rtsp-stream-to-virtual-web-camera

Hourly Range: $10.00-$20.00

Posted On: January 20, 2024 21:12 UTC
Category: Full Stack Development
Skills:Python, Video Stream, WebRTC

Country: France

click to apply