Research on the Identification Method of Unsafe Behavior of Construction Workers Based on Deep Learning

This paper focuses on exploring a deep learning-based approach to identify and address unsafe behaviors exhibited by construction workers. The proposed method utilizes a deep convolutional neural network (CNN) to analyze surveillance videos captured at construction sites, with the objective of automatically detecting and alerting unsafe behaviors to enhance overall site safety. Through extensive experimentation, it is observed that this approach achieves remarkable accuracy and real-time performance, thus effectively mitigating safety risks associated with construction sites. By utilizing deep learning techniques and specifically the deep convolutional neural network (CNN), the proposed method enables the automatic detection and warning of unsafe behaviors exhibited by construction workers. This real-time identification of hazards can greatly contribute to minimizing accidents and improving overall safety conditions at construction sites. Implementing this method in practice could lead to a substantial reduction in safety risks, thereby safeguarding the well-being of workers, and enhancing the overall safety culture within the construction industry.