Safety and Compliance Management System Using Computer Vision and Deep Learning

With the advancement in the Deep Learning and Computer Vision-based methodologies, the safety and productivity analysis of the construction entities (workers, equipment, and structures) on-site can potentially be improved. Construction industry is majorly the most dangerous sectors with respect to occupational safety and health because of the chaotic and dynamic environment at construction sites. The study of safety and compliance monitoring measures at construction by using the computer vision and deep learning-based approaches is the key focus of this paper. Safety and compliance are acclamatory to each other and are very essential for managing the safe conditions at construction. Manual monitoring of the compliant and unsafe conditions at the site is a challenging job and thus, a smart framework for automated monitoring is necessary for the assurance of a safe and healthy environment. A framework for safety and compliance management system is presented in this paper for enhancing the safer work culture along with the measures. This paper examines the prior literature on computer vision as per the safety and compliance constraints for (I) Understanding the existing SOTA (state-of-the-art) methods and their respective outcomes (II) Finding the challenges and limitations in the currently employed approaches and (III) Providing the potential directions for a future line of work. The purpose is to build a culture of safety using Computer Vision and Deep Learning.