سيدسياوش سرمدراد

Timely identification and monitoring of workers' use of personal protective equipment (PPE) can effectively reduce the risk of accidents in the construction industry. Continuous supervision of PPE usage not only enhances workplace safety but also ensures the health and security of workers. Accurate and timely assessment of PPE usage prevents accidents and reduces costs associated with medical treatments and damage caused by such incidents. The use of advanced technologies to monitor PPE improves the efficiency and effectiveness of these protective measures. This approach enables better protection of workers' health and safety while significantly reducing workplace accidents. Moreover, continuous and precise eva‎luation of protective equipment fosters a safety culture in the workplace and increases workers' motivation to comply with safety regulations. Most research on this subject has primarily focused on detecting the presence of workers and determining whether or not they are wearing helmets. These studies have largely been limited to simple detection of PPE, with their results serving mainly research purposes. The outputs typically consist of rectangular bounding boxes and basic labeling, offering limited practical applications. To overcome these limitations, this study leverages deep learning and computer vision techniques to assess and mitigate safety risks for construction workers. The advanced YOLOv8 algorithm, a state-of-the-art model for object detection, was selected for this purpose. To train and eva‎luate this model, a dataset comprising images of work environments and workers' PPE in various conditions from construction sites in Iran (domestic) and other countries (international) was collected and utilized. After training the PPE detection model, a risk assessment index for PPE non-compliance was calculated using the results of other studies, enabling workplace safety levels to be classified. The findings of this study demonstrate that deep learning and computer vision techniques can significantly reduce safety risks in construction environments. It is hoped that the development and widespread adoption of such systems will enhance safety levels in the construction industry and create safer workplaces for workers.

صنعت ساخت , مديريت ايمني , مديريت ريسك , بينايي رايانه , يادگيري عميق

Construction Industry , Safety Management , Risk Management , Computer Vision , Deep Learning

Seyed Siavash Sarmad Raad

Dr.Parviz Ghoddusi-Dr.Hassan MalekiTabar