Review on the Evolution of Civil Structure Health Monitoring Technology from Physical Detection to Data Intelligence

Abstract

Civil engineering structures face various service risks, and the drawbacks of traditional manual testing are becoming increasingly prominent, driving the evolution of structural health monitoring from physical testing to data intelligence. This article systematically reviews the development of structural health monitoring technology, elucidates its core connotations and the evolution of damage identification tasks, and elaborates on the technical principles and applications of traditional non-destructive testing techniques in structural damage detection. It focuses on the application of machine learning in structural health monitoring, as well as technological breakthroughs in pixel level damage localization, fatigue crack dynamic capture, 3D damage assessment, and unlabeled dynamic response measurement using computer vision, and their engineering practices. Finally, the key challenges of current structural health monitoring technology were summarized, including poor algorithm robustness in extreme environments, insufficient cross scenario generalization ability, and lack of full lifecycle life prediction capability. This article provides a reference for the intelligent operation and maintenance of civil structures.