Personalized Edge-Cloud Collaborative High-Fall-risk Monitoring Strategy for Grid Workers at Heights

Falling from height is one of the most frequent causes of death for power grid workers working at heights. It is necessary to conduct reasonable high-fall risk detection for power workers working at heights. However, simply using sensor data from nearby workers for safety assessment is highly inaccurate, and using high computing power data centers far from workers for accurate safety analysis can easily lead to untimely warnings. This paper proposes a collaborative edge-cloud high-fall risk detection strategy based on worker posture and multiple sensors. This can solve the delay problem and ensure monitoring accuracy. Considering that the work posture of workers is highly personalized. Therefore, this paper designed a personalized risk assessment system for high-fall based on the Gaussian model (GM) and the Gaussian Mixture model (GMM). Finally, the effectiveness of the proposed strategy was verified through video frames of workers working at heights, and the accuracy of the personalized pose risk identification model established was verified.