Welding Workers Without Protective Gear Violation Detection Dataset

In the current industrial field, welding operations pose significant safety risks, especially when workers do not wear necessary protective equipment, which can easily lead to eye injuries. However, existing safety monitoring systems mostly rely on manual inspection, which is inefficient and cannot provide real-time feedback. To effectively improve worker safety, this dataset aims to automatically recognize whether workers are wearing protective gear during the welding process using video monitoring technology and send alerts promptly. Data are sourced from multiple industrial welding sites, recorded around the clock using high-quality camera equipment to ensure the clarity and completeness of the videos. The data processing involves multiple rounds of labeling and expert review mechanisms to ensure consistency and accuracy in annotations. Data will be stored in MP4 format for subsequent deep learning model training and analysis.