Examination of Risky Behvior Detection in Crane Operation Work

Abstract

In recent years, the IoT (Internet of Things) has been attracting attention as a basic technology that supports Society 5.0. We are promoting research and development to ensure product quality and worker safety in the industrial field by utilizing IoT. The purpose of this research is to automatically detect dangerous behavior of workers in crane operation work from a camera at a manufacturing site that handles large products. In this study, we conducted an experiment to detect dangerous behavior for four items of crane operation work using video data obtained from on-site surveillance cameras. It was confirmed that the positions of the worker and the suspended load could be sensed by person detection and object detection, and dangerous acts could be detected from their positional relationship.