Computer Vision for Bridge Inspection and Monitoring

Bridge inspection mostly rely on visual assessment in practice. Therefore, digital image-based structural inspection has been proposed as early as in 1980s and become the most fruitful computer-science aided field in Civil Engineering. The approach, termed either computer-vision or machine-vision method, starts with the use of digital cameras or other imaging devices, then an image analysis or understanding algorithm is adopted to detect, classify, and even localize structural surface damage. Moreover, by collecting continuous imagery data in the spatial or temporal domain, vision methods can be adapted to measure structural motions in terms of local deformation or structural vibration. This chapter introduces the computer vision based inspection and monitoring technologies for assessing bridge structures. Specifically, this chapter introduces basic concepts and history of computer vision, conventional methods, deep learning methods, and applications for bridge damage inspection and structural health monitoring with real examples.