In the afternoon of May 8, at the invitation of Prof. Ying Luo, Director of International Joint Research Center for Health Management of Key Structures of High-end Equipment of the Ministry of Science and Technology of the University, Prof. Fuh-Gwo Yuan of North Carolina State University gave a lecture titled "Advances in Machine Learning & Computer Vison in Structural Health Monitoring" in the Third Lecture Hall of the University Conference Center. . The lecture was hosted by Prof. Zhu Jianguo, Director of Structural Health Management Institute of the university, and more than 100 teachers and students from the School of Geotechnical Engineering, the School of Physics and Electronic Engineering, and the School of Automotive and Transportation Engineering attended the lecture.

Prof. Fuh-Gwo Yuan is the Foreign Director of our National Cooperation Base, Professor/BoD of the School of Mechanical and Aerospace Engineering, North Carolina State University, and Director of the Laboratory of Intelligent Structures and Materials in the Department of Mechanical and Aerospace, USA, with research expertise in structural health monitoring, damage durability of composite structures, intelligent materials and structures, and fracture and life prediction of advanced materials and structures, and has published high-level He has published more than 150 academic papers, and was awarded the International SHM Person of the Year Award in 2013, and the Outstanding Contribution Award in the field of NDE by the SPIE Society in 2023.

This presentation summarizes the research results of Prof. Yuan's group on the application of computer vision combined with deep learning in structural health monitoring (SHM) in recent years.Prof. Yuan firstly introduces the current application of SHM in practice and the challenges it faces. Then, Prof. Yuan details the applications of deep learning algorithms in breaking the bottleneck of SHM, which include: visual inspection of corrosion and fatigue cracks in structures using augmented virtual reality (AR) devices; localization of impact locations in structures using deep convolutional neural networks (CNNs) and recurrent neural networks (RNNs); damage imaging using high-speed cameras, as well as the use of 3D DIC and off-axis 2D DIC for damage identification. Finally, Prof. Yuan emphasized that with the continuous development of optical and computer technologies, the application of deep learning methods in the field of SHM will become more widespread.

In the final interactive session, the participating students and faculty members actively asked questions based on the actual scientific research and the content of the report, and Prof. Yuan also gave detailed answers based on his own scientific research experience and practical experience. This academic exchange has broadened the international scientific research and academic vision of teachers and students, and provided new directions and ideas for their subsequent scientific research work.