CCTV_SENTRY_YOLO12

Overview

CCTV_SENTRY_YOLO12 is an advanced object detection system built using YOLOv12 by Ultralytics. It provides real-time monitoring, object tracking, and line-crossing detection for IP camera streams. Hosted on Hugging Face Spaces, it enables users to easily interact with the model via a web interface.

Features

• Line-Crossing Detection: Detects objects crossing user-defined lines.
• Real-Time Object Detection: Utilizes YOLOv12 for high-speed object tracking.
• Interactive Interface: Powered by Gradio for an intuitive user experience.
• Customizable Classes: Filter objects based on specific detection classes.
• Detailed Visualization: Annotated frames with bounding boxes, IDs, and counts.

demo.mp4

Model Details

The system leverages the YOLOv12 model from Ultralytics for accurate and efficient object detection. Key technologies include:

• OpenCV: For video frame processing.
• Gradio: For creating an interactive user interface.
• Ultralytics YOLO: For state-of-the-art object detection and tracking.

How It Works

1. Upload or provide the URL of an IP camera stream.
2. Draw a line on the first frame to set the detection boundary.
3. Select the object classes to monitor.
4. Watch real-time detections and line-crossing counts directly on the interface.

Industrial and Commercial Applications

Parking Management

• Vehicle entry/exit tracking
• Parking space occupancy monitoring
• Unauthorized parking detection

Manufacturing

• Conveyor belt product counting
• Quality control inspections
• Real-time inventory tracking

Retail and Logistics

• Customer movement analysis
• Stock level monitoring
• Theft prevention systems

Transportation

• Vehicle tracking
• Loading dock management
• Traffic flow analysis

Security

• Perimeter surveillance
• Restricted area monitoring
• Crowd density estimation

Usage

Requirements

• Python 3.x
• Install dependencies:

  pip install ultralytics opencv-python-headless gradio numpy pillow

Run Locally

1. Clone the repository:

   git clone /~https://github.com/SanshruthR/CCTV_SENTRY_YOLO12.git
   

2. Navigate to the project directory:

   cd CCTV_SENTRY_YOLO12
   pip install -r requirements.txt

3. Start the application:

   python app.py

Live Demo

Experience the project live on Hugging Face Spaces:
CCTV_SENTRY_YOLO12 on Hugging Face

Deployment

• To create a live HLS stream for testing, refer to this GitHub repository:
  /~https://github.com/SanshruthR/mock-hls-server

• Use the sample video file for testing:
  https://videos.pexels.com/video-files/1169852/1169852-hd_1920_1080_30fps.mp4

License

This project is licensed under the MIT License.