Abstract:

To address the limitations of single types of devices (e.g., sensors, intelligent cameras) in perceiving safety risks during the high-risk command-operation interactions in lifting operations, an intelligent device integrating computer vision technology and sensors (easily called an intelligent device) was developed for the collaborative perception of safety risks in lifting operations. First, the motion characteristics of the signalman and the crane in lifting were analyzed. Then, to eliminate the risk of crane operators' misoperation in lifting, a risk cooperative perception framework for the intelligent device was proposed in combination with the different motion characteristics and work requirements of man-machine. Finally, the feasibility test of the risk perception accuracy and risk perception latency of intelligent device was carried out in the laboratory simulation environment. The lab-based simulated experiment results demonstrate that the intelligent device could collaboratively perceive the safety risks during command-operation interaction in lifting operations. Instances of misoperation by crane operators can be alarmed in real-time, with a risk perception accuracy of 95.17% and a risk perception latency of approximately 0.25 s.

Key words: lifting operations, risk perception, collaborative perception, intelligent device, mixed convolutional neural network-lightweight (MCN-Lite)