Cloud-based database framework of corrosion detecting and grading for temporary steel pipe supports

Failures of temporary steel pipe supports due to corrosion present significant safety risks and economic challenges within the construction industry, stemming from the absence of robust logging and quality inspection systems. This paper introduces a novel cloud-based database framework designed for the automated detection and grading of these critical components, aiming to enhance site safety and equipment management. The proposed multi-detection system utilizes state-of-the-art deep learning architectures, including YOLOv9 for object detection and YOLOv9-seg for instance segmentation. On held-out test sets, the system demonstrated robust performance, achieving a mean average precision (mAP@[.5:.95]) of 0.71 for the precise detection of steel pipe supports. These analytical capabilities are integrated into a unified web server, which combines web development technologies with equipment analysis results. This platform facilitates real-time data visualization and employs a novel, practical checklist-based grading system to systematically assess equipment condition according to industry standards and site-specific repairability. The primary contribution is a comprehensive online management system that provides a data-driven guide for quality assurance, enabling timely interventions and reducing accidents caused by faulty temporary equipment. Furthermore, the framework’s adaptable design demonstrates strong potential for replicability across diverse construction sites and for other types of defects, paving the way for more effective, technology-driven safety protocols.