Object Detection Using YOLOv5

This repository contains a Python script for performing object detection on images using the YOLOv5 model. The detected objects are saved to a PostgreSQL database, and the results can be visualized using Matplotlib. The project is designed to process images from a specified directory, detect objects using YOLOv5, and store the results in a structured database for further analysis.

Table of Contents

1. Features
2. Requirements
3. Installation
4. Usage
5. Database Schema
6. Visualization
7. Logs and Monitoring
8. Next Steps
9. Contributing
10. License

Features

• Object Detection: Uses YOLOv5 to detect objects in images.
• Database Integration: Saves detection results (class label, confidence, bounding box coordinates) to a PostgreSQL database.
• Visualization: Visualizes detected objects with bounding boxes and labels on the images.
• Logging: Detailed logging for monitoring and debugging.
• Modular Code: Well-structured and modular code for easy maintenance and extension.

Requirements

• Python 3.8 or higher
• PostgreSQL database
• Required Python libraries:
  • opencv-python
  • torch
  • torchvision
  • sqlalchemy
  • matplotlib

Installation

1. Clone the Repository: bash git clone https://github.com/Azazh/Object-Detection-Using-YOLO5.git cd object-detection-yolo

2. Set Up a Virtual Environment: bash python -m venv venv source venv/bin/activate # On Windows, use `venv\Scripts\activate`

3. Install Dependencies: bash pip install -r requirements.txt

4. Set Up PostgreSQL Database:

   • Create a database named medical_dw (or any name you prefer).
   • Update the DB_CONNECTION string in the script with your database credentials: python DB_CONNECTION = "postgresql://username:password@localhost:5432/medical_dw"

5. Download YOLOv5 Model: The script automatically downloads the YOLOv5 model using torch.hub. Ensure you have an active internet connection during the first run.

Usage

1. Prepare Images:

   • Place your images in the ../raw_data/media directory (or update the IMAGE_DIR variable in the script).

2. Run the object_detection script: bash python object_detection.py

3. View Results:

   • Detection results are saved in the detection_results table in the PostgreSQL database.
   • A sample visualization of the first image is displayed using Matplotlib.

Database Schema

The detection results are stored in the detection_results table with the following schema:

| Column | Type | Description | |||--| | id | Integer | Primary key (auto-increment). | | image_path | String | Path to the image file. | | class_label | String | Detected object class label. | | confidence | Float | Confidence score of the detection. | | x_min | Integer | Bounding box top-left x-coordinate. | | y_min | Integer | Bounding box top-left y-coordinate. | | x_max | Integer | Bounding box bottom-right x-coordinate. | | y_max | Integer | Bounding box bottom-right y-coordinate. |

Visualization

The script includes a function to visualize the detected objects on the images. It draws bounding boxes and labels (with confidence scores) on the images using Matplotlib.

Example:

Logs and Monitoring

The script logs all activities to object_detection.log. You can monitor the logs for errors, warnings, and informational messages.

Example log: 2023-10-10 12:34:56,789 - INFO - Found 100 images for processing. 2023-10-10 12:35:10,123 - INFO - Detected 3 objects in ../raw_data/media/image1.jpg. 2023-10-10 12:35:15,456 - INFO - Saved 3 detections to database.

Next Steps

• Fine-tune YOLO: Train the YOLOv5 model on custom datasets for better accuracy.
• Extend Visualization: Add support for visualizing multiple images and saving visualizations to disk.
• Integrate with Data Warehouse: Combine detection results with other data sources for comprehensive analysis.
• Add Unit Tests: Write unit tests for critical functions like detect_objects and save_to_db.

Contributing

Contributions are welcome! If you'd like to contribute, please follow these steps: 1. Fork the repository. 2. Create a new branch for your feature or bugfix. 3. Commit your changes. 4. Submit a pull request.

License

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

Acknowledgments

• YOLOv5 by Ultralytics for the object detection model.
• SQLAlchemy for database operations.
• OpenCV for image processing.
• Matplotlib for visualization.