Abstract: Development of computer vision technology is important to structural health monitoring since deflection is a key factor to large structures. In this study, a monocular video deflectometer has been constructed by developing an efficient fast algorithm based on Digital Image Correlation. The precision of the video deflectometer has been verified in the outdoor experiment. This system has been adopted to dynamically measure the vertical deflection of the Yingwuzhou Yangtze River Bridge at given positions. It is found that the measurement is in good agreement with the readouts from the existing GPS sensors which are embedded in the bridge.