Computer Vision-Based Three-dimensional (3D) Vibration Displacement Measurement for Civil Structures

Abstract

This thesis proposes a novel computer vision-based target-free full-field 3D displacement measurement for civil engineering. Two cameras are combined as a binocular camera system to acquire the vibration video of civil structures. Some state-of-the-art methods including deep learning in the computer vision area are applied to realize target-free, full-field and tiny displacement measurements. A series of experimental tests and an in-field vibration test are conducted to validate the proposed vision measurement system.