Smart Dust-Free Protective Equipment and Cleanroom Inspection System

Abstract

A B S T R A C T

In response to the stringent safety requirements of semiconductor cleanrooms, this study aims to develop an advanced dust-free protective equipment inspection system tailored for personnel entering these critical environments. Central to this system is the application of human pose detection techniques, which precisely identify essential body parts such as the head, body, hands, and foot. These detections serve as the foundation for deep learning network architectures that rigorously evaluate the adequacy of protective suits worn by personnel. Additionally, a system is designed to monitor the effectiveness of dust removal and ensure comprehensive coverage during the air shower process, crucial for maintaining impeccable cleanliness standards. The efficacy of the proposed pipeline is substantiated through rigorous validation encompassing comprehensive quantitative and qualitative analyses. Initial trials demonstrate robust perfor- mance, with quantitative accuracy rates of 98.42% for the protective equipment inspection system and an error rate of approximately 5% for the air showering process system. These results affirm the system’s capability to reliably assess adherence to safety protocols in real-time scenarios. Beyond its immediate application in semiconductor cleanrooms, this adaptable system holds promise for integration into diverse sectors where stringent safety and contamination- free environments are paramount. Future research aims to enhance the system’s adaptability to varying operational conditions and expand its functionalities through advancements in real-time feedback mechanisms and integration with edge computing technologies.