Computer Vision-Enabled Safety for Construction: A Deep Learning and Predictive Analytics Approach Across Project
Authors: Sai Kothapalli
DOI: https://doi.org/10.37082/IJIRMPS.v12.i4.232598
Short DOI: https://doi.org/
Country: USA
Full-text Research PDF File:
View |
Download
Abstract: Construction sites remain among the most hazardous work environments, with injury rates significantly exceeding other industries. This research presents a comprehensive analysis of machine learning (ML) applications for improving safety management in construction environments, with a specific focus on infrastructure projects in Austin, Texas. Through implementation of computer vision-based personal protective equipment (PPE) detection, predictive analytics for accident prevention, and real-time hazard identification systems, construction sites demonstrated a 34% reduction in safety incidents over a 12-month period. The study analyzes data from five major construction projects totaling $2.3 billion in infrastructure investment, including the Austin-Bergstrom International Airport expansion and downtown high-rise developments. Key findings indicate that ML-powered safety systems achieve 92.7% accuracy in PPE compliance detection and 87.3% precision in predicting high-risk scenarios. This research contributes to the growing body of knowledge on smart construction technologies and provides empirical evidence for the effectiveness of ML-driven safety interventions.
Keywords:
Paper Id: 232598
Published On: 2024-08-09
Published In: Volume 12, Issue 4, July-August 2024