PPE Detection System

01 Overview: A safety check that runs on a camera feed

The PPE Detection System uses a YOLOv8 object-detection model with OpenCV to recognise personal protective equipment (helmets, safety vests and masks) in real time.

When required equipment isn't detected, the system generates a PPE violation alert. A Flask application serves the web interface, built with HTML and CSS.

02 Problem: Compliance is easy to miss when it's checked by eye

Protective equipment only works if it's actually worn, and watching for it manually is repetitive work that's easy to lose track of.

The project explores how a detection model can take the first pass automatically and surface only the moments that need a person's attention.

03 What I Built: Detection, alerts and a web interface

• Real-time detection of helmets, safety vests and masks with YOLOv8.
• Image and video-frame handling with OpenCV.
• PPE violation alerts, generated when required equipment is missing.
• A Flask web application with an HTML and CSS interface for running and viewing the system.

04 Technical Approach: From frame to alert

1. 01CaptureOpenCV reads frames from the video input.
2. 02DetectYOLOv8 locates helmets, vests and masks in each frame.
3. 03CheckDetections are compared against the required equipment.
4. 04AlertMissing PPE is raised as a violation alert.
5. 05ServeFlask delivers the results to the web interface.

Built with

• Python
• YOLOv8
• OpenCV
• Flask
• HTML5
• CSS3