Deploying Computer-Based Vision to Enhance Safety in Industrial Environment

One of the cornerstones of Lean manufacturing, according to academic research, is the 5S+1, which introduces quality in manufacturing. This study aims to demonstrate how computer-based vision and object detection algorithms, can assist in the implementation of safety as 6th S in 5S+1 by monitoring and identifying employees who disregard accepted safety procedures, like wearing Personal Protective Equipment (PPE). The research evaluated the performance indicators of a detection technique and reviewed and analyzed it. To confirm workers’ PPE compliance, the suggested model used the You-Only-Look-Once (YOLO v7) architecture. A deep learning technique was subsequently applied to confirm the safety helmets and safety vests. This strategy is determined to be the most effective when using the VGG-16 algorithm, achieving an 80% F1 score and processing 11.79 frames per second (FPS), making it ideal for real-time detection.