Using existing CCTV network for crowd management using AI&ML.

This is a real-time Object and Face Detection System using Python, OpenCV, YOLOv3 for object detection, and Haar Cascade for face detection. It detects and tracks objects and faces in live video streams from a camera source.

• Object detection using YOLOv3.
• Face detection using Haar Cascade.
• Real-time video feed from a camera source.
• Database integration to store detection results.
• Unique ID assignment for detected faces.
• Frame rate calculation.

Before running this project, you need to have the following:

• Python 3.x installed.
• OpenCV Python library (cv2) installed.
• SQLite3 for Python installed.
• YOLOv3 weights, configuration file, and coco.names file (for object detection) in the project directory.
• Haar Cascade XML file for face detection.

1. Clone this repository to your local machine:

   (https://github.com/i-soumya18/Garud-Drishti)
   
   

cd object-face-detection

python main.py

#Press 'q' to exit the application.

#The project uses an SQLite database (detection_database.db) to store detection results. It has two tables:

-object_detections: Stores information about detected objects. -face_detections: Stores information about detected faces, including their unique IDs.

You can customize this project by:

-Changing the YOLOv3 model weights and configuration files for different object detection tasks. -Modifying the Haar Cascade XML file for different face detection tasks. -Adjusting the confidence threshold for object detection. -Fine-tuning detection parameters according to your requirements.

If you'd like to contribute to this project, please follow these steps:

-Fork the repository on GitHub. -Clone your forked repository to your local machine. -Create a new branch for your feature or bug fix. -Make your changes and commit them. -Push your changes to your forked repository. -Submit a pull request to the original repository.

This project is licensed under the MIT License - see the LICENSE file for details.

-This project uses the YOLOv3 model for object detection. For more information about YOLO, visit: YOLO: Real-Time Object Detection. -Face detection is based on Haar Cascade Classifiers provided by OpenCV.

If you have any questions or suggestions, please feel free to contact the author.