Enhancing Construction Safety: A Computer Vision Approach for Proactive Near-Miss Events Identification

Construction projects are prone to high-risk safety accidents, highlighting the importance of effectively managing near-miss events to prevent such incidents. However, the reliance on manual supervision in traditional monitoring and early warning systems makes timely detection of near-miss events challenging. This study presents an integrated approach that combines hazard energy source analysis with computer vision technology to identify near-miss accidents on construction sites. The hazard energy source method provides a foundational framework for identifying near-miss accidents, while computer vision technology enables active detection of such incidents. The integration of these methods facilitates proactive accident prevention. To assess the efficacy of the proposed approach, an analysis of hazard energies and their coupling mechanisms is conducted, alongside the verification of the computer vision model’s performance. The proposed approach offers a promising tool for enhancing safety management on construction sites.