Computer-vision methods for structural monitoring have attracted much attention in recent years thanks to the possibility of measuring an unprecedented number of displacement time histories from a single sensor in the absence of stationary points close to the structure, using hardware with accessible costs that is simple to set up and operate by structural engineers with basic video photography knowledge and programming skills. Past experimental studies showed the potentialities of computer-vision algorithms to provide accurate results that can be equivalent to those obtained from accelerometers and displacement transducers, even when very small displacements are experienced, as in the case of stiff structures and/or low load levels typical of operational conditions. Such achievement is made possible thanks to the development in the last decade of sophisticated and computationally efficient computer-vision algorithms able to achieve resolutions that are a small fraction of the dimension of a single pixel in the captured video footages, while being able to track large displacements. This opens unexplored territories in structural identification and condition assessment of civil structures, freeing structural engineers from the inevitable limitations imposed by a reduced number of pointwise contact sensors, as is commonly the case in many temporary or permanent structural monitoring setups. Accordingly, the objective of this article is to provide a critical discussion of the potential benefits of computer-vision structural monitoring applied to a full-scale building where the use of traditional contact sensors faces major difficulties and limitations. Insight is given to the extraction of the time-history of displacements, with compensation of vibrations and external disturbances in the video camera.