How to Use Your Existing Cameras for AI (Fast, Easy, No New Hardware)

Think you need brand-new cameras to run modern computer vision? Think again. Whether it’s a CCTV setup, a workshop camera, or even your smartphone, your current hardware is all you need.

All you need to do is stream video frames from your existing camera into a Roboflow Workflow (via RTSP or similar protocols), then apply any of the latest computer vision models that Roboflow offers.

---

Usage

1. Clone this repository:

   git clone https://github.com/Neurl-LLC/roboflow-04.git
   cd roboflow-04

2. Create a workflow on Roboflow, or simply clone this example workflow.

3. Create a .env file with the following environment variables:

   ROBOFLOW_API_KEY="YOUR_ROBOFLOW_API_KEY"
   WORKSPACE_NAME="YOUR_WORKSPACE_NAME"
   WORKFLOW_ID="YOUR_WORKFLOW_ID"
   VIDEO_REFERENCE="YOUR_RTSP_LINK"

4. Install dependencies:

   pip install -r requirements.txt

5. Run the producer script (handles inference):

   python producer.py

6. Run the consumer script (handles display):

   python consumer.py

---

Simulating an IP Camera

Don’t have an IP camera handy? No problem. You can simulate one using MediaMTX and FFmpeg.

1. Install MediaMTX and download FFmpeg.

2. Start the MediaMTX server:

   ./mediamtx

3. Stream a sample video to the MediaMTX server with FFmpeg:

   ffmpeg -re -stream_loop -1 -i Street_Walk.mp4 -c copy -f rtsp rtsp://127.0.0.1:8554/mystream

This makes the sample video available at:

rtsp://127.0.0.1:8554/mystream

Use this RTSP link as your VIDEO_REFERENCE in the .env file.

---

✅ That’s it. You now have AI-powered computer vision running on your existing cameras with no new hardware required.