Preventing Drowsy Driving Accidents in the Construction Industry Using Computer Vision and Convolutional Neural Networks

Construction machine operators are at risk of drowsy driving, which can lead to accidents and injuries on jobsites. Jobsite injuries are a major concern in the construction industry, with operators of heavy machinery being particularly vulnerable. According to the US Bureau of Labor Statistics, in 2019, the fatal injury rate for workers in the construction industry was higher than the national average for all workers, and many of these fatal injuries were caused by heavy equipment. To address this issue, a method using computer vision and edge devices has been proposed to detect drowsy construction machine drivers using convolutional neural networks (CNNs). A dataset of images showing signs of drowsiness was constructed and used to train the CNN. After training, the neural network can score the testing images and predict if the subject in the image is drowsy or non-drowsy. This technology has the potential to greatly improve safety in the construction industry by alerting drowsy operators before an accident occurs, thus reducing the risk of accidents and injuries on jobsites. The implementation of this technology can help in reducing jobsite injuries and improve the overall safety of construction sites.