PPE Detection - Personal Protective Equipment Detection using YOLOv8 Object Detection

Introduction

PPE Detection is a Python-based project that employs YOLOv8, a robust object detection model, to identify and classify personal protective equipment (PPE) on construction sites. The project involves downloading a YOLOv8 model from Roboflow, training it on Google Colab, and using the best.pt model to detect PPE in images or videos.

Prerequisites

• Python 3.6 or higher

• Virtualenv (optional but recommended)

Installation

1. Create a virtual environment and activate it:

virtualenv venv
source venv/bin/activate    # On Windows: venv\Scripts\activate

2. Install required packages:

pip install -r requirements.txt

Training the Model

1. Download YOLOv8 model from Roboflow:

• Download the YOLOv8 model file from Roboflow and place it in the weights/ directory.

2. Train the model on Google Colab:

• Follow the steps outlined in the colab file (Yolov8.ipynb) to train the model using your custom dataset.

Usage

1. Prepare your video:

• Place your video file in the project directory.

2. Run the PPE Detection:

python PPE-Detection.py 

3. View the results:

• The processed video with PPE Detections will identifies and labels present equipment (e.g., "mask") and denotes absence as "no-mask" for each equipment type, providing accurate results.

Configuration

• You can customize the confidence threshold for PPE detection by modifying the conf parameter in PPE-Detection.py. The default value is set to 0.5.

Acknowledgements

• Ultralytics
• CVzone library by CVzone

Contribution

Contributions are always welcome!

If you find any issues or have suggestions for improvements, feel free to create a pull request.