基于深度学习的土木工程计算机视觉健康监测

Deep Learning-Based Computer Vision for Health Monitoring in Civil Engineering

土木工程领域的健康监测对保证工程长期、稳定服务有着重要的意义。相较于传统的监测方法,基于深度学习的计算机视觉技术具有高效、准确等优势。对基于深度学习的计算机视觉技术在土木工程全生命周期健康监测领域中的应用进行系统综述。首先,借助文献可视化软件对该领域文献进行科学计量分析;其次,简要阐述了计算机视觉技术的发展历程,总结了在构建深度学习数据集过程中数据获取、数据处理和数据标注三个重要环节的方法与内容;最后,重点回顾了在施工现场安全管理、在役结构局部损伤检测和结构灾后整体损伤评估等应用场景中基于深度学习的计算机视觉技术的发展历程与工程实际应用价值,并展望了可拓展的应用方向。

Health monitoring in the field of civil engineering is of great significance to ensure the long-term and stable service of infrastructure.Compared with traditional monitoring methods,the computer vision technology based on deep learning has the advantages of high efficiency and accuracy.This paper provides a systematic review on the application of the deep learningbased computer vision technology in the field of civil engineering life cycle health monitoring.First,a scientific econometric analysis of the literature in this field is conducted with the help of literature visualization software.Then,the development process of computer vision technology is briefly described,and the methods of data acquisition,data processing,and data annotation in the process of constructing deep learning data sets are summarized.Afterwards,the development and practical engineering application value of the computer vision technology based on deep learning in safety management of construction site,local damage detection of in-service structures and overall damage assessment of structures after disaster are reviewed.Finally,the future application directions are prospected.