Development of a pose estimation-based detection system for point and call confirmations at pedestrian crossings

Abstract

Safety measures such as pointing and checking when crossing pedestrian crossings have been implemented in the manufacturing industry to prevent accidents involving falls when walking in factories and to raise safety awareness. However, because there is no system to judge the implementation status of pointing and checking, there is a problem that it has not been sufficiently established among all employees. In this study, we propose a pointing confirmation judgment system at pedestrian crossings with a feedback function. Based on the skeletal information acquired using posture estimation AI, we developed a system that judges whether or not a person crosses a pedestrian crossing by pointing and checking, and verified its operation.