AI-Driven Multipoint Harness Inspection System

Falls from a height are still one of the major sources of accidents on construction, mining, and industrial work sites. Despite safety codes requiring harness use, helmet use, and fall protection equipment, non-compliance and misuse remain outstanding hazards. The Multipoint Harness Inspection System is a low-cost, portable, AI-based safety monitoring system introduced in this paper that is meant to provide real-time compliance with harness use. The system incorporates computer vision through a TensorFlow Lite image classification model, IoT hardware elements such as a buzzer, battery pack, buck converter, and battery management system, and edge computing on a Raspberry Pi 5. In contrast to cloud-based monitoring systems, the device is completely offline and has real-time detection and instant alerting at building construction sites. Through the integration of AI-powered harness detection with an ultra-low-power embedded platform, the system improves workplace safety.