• Version
• Download 14
• File Size 677.28 KB
• File Count 1
• Create Date 20 March 2026
• Last Updated 20 March 2026

Safevision:Automated PPE Detection and Violation Monitoring using Yolov8 and Deepsort

Krishnendu Patni, Sachin Kumar, Swati Dwivedi, Nitin Kumar
krishnendupatni5@gmail.com
Department of BCA (Batch: 2023–2026), Haridwar University, Roorkee, Haridwar
Internal Guide: Mr. Gaurav Kumar, Asst. Professor -nayakgaurav682@gmail.com

Abstract:Industrial workplaces such as construction sites, chemical plants, steel industries, and warehouses face significant safety challenges due to non-compliance with Personal Protective Equipment (PPE) regulations. Manual supervision methods are inefficient and prone to human error. This paper presents SafeVision, a real-time Artificial Intelligence (AI)-based PPE detection and safety monitoring system designed to automate compliance verification using computer vision and deep learning. The system integrates YOLO-based object detection, DeepSORT-based multi-object tracking, and a web based dashboard built with modern web technologies. SafeVision detects safety violations, logs events in a centralized database, triggers real-time alerts, and provides actionable analytics for safety administrators. Experimental evaluation demonstrates high detection accuracy with real-time processing capabilities, making the system suitable for industrial deployment. The proposed solution contributes toward Industry 4.0 by enhancing automated workplace safety monitoring.
Keywords:Artificial Intelligence, Computer Vision, PPE Detection, YOLO, DeepSORT, Workplace Safety, Real-Time Monitoring,Industrial Automation