Method for determining personal protective equipment compliance using computer vision

Abstract

A method for determining personal protective equipment compliance using computer vision comprising a desktop computer system unit, configured to receive real-time video frames from a camera member, execute a PPE classifier module, and coordinate a trained object-detection deep learning model; a camera member configured to capture real-time video frames; a trained object-detection deep learning model configured to detect persons and personal protective equipment items and generate corresponding person and PPE bounding boxes; a PPE classifier module configured to define predefined body zones within each person bounding box and determine compliance by computing spatial overlap between each PPE bounding box and a corresponding body zone against a stored threshold; and a monitor member configured to display processed video frames with visually distinguished compliance status. The utility model enables automated real-time detection of individuals and personal protective equipment and objective compliance classification based on spatial analysis of detected objects.