Enhancing Industrial Safety: A Comprehensive Approach to Detecting and Mitigating Potential Hazards

Industrial safety has always been one of the important factors that restrict the sustainable development of enterprises. These safety hazards threaten the lives and health of workers at all times. Therefore, improving industrial safety levels is a necessary condition for the sustainable development of enterprises. Currently, most detection devices only address single problems, and these single-task devices cannot meet the monitoring needs of multiple safety hazards. This study uses object detection technology to achieve real-time monitoring and identification of safety helmets, fires, and industrial instruments, which can simultaneously solve multiple safety hazards and improve industrial safety levels. Specifically, we employ the PP-YOLOE algorithm for multi-task target detection, enabling real-time detection of these three safety hazards with recognition precision reaching 51.2%. Furthermore, building upon target detection, we employ image segmentation and computer vision techniques to achieve accurate readings of industrial instruments, achieving a precision of up to four decimal places. Finally, the system is successfully deployed on a quadruped robot, coupled with autonomous navigation and path planning techniques, facilitating robot inspections in industrial environments. Extensive experiments demonstrate the effectiveness of this approach in enhancing industrial safety levels and ensuring the sustainable development of enterprises.