In image-based displacement measurement, the distance from a camera to the object directly affects the accuracy of the measurement, which limits the range of objects for which displacement can be estimated with the use of a single video camera with a wide-angle lens. In this study, we develop a method to integrate the data for the displacement estimated by multiple video cameras. Our method utilizes overlapping measurement points, with the aim of estimating displacements over a wide area at a time. The proposed method first aligns displacement waveforms to minimize the squared error of the displacement time series at overlapping measurement points, which enables frame-by-frame time synchronization. Next, the true state at each location is estimated with consideration for time synchronization errors and displacement estimation errors using a multivariate state-space model, which considers cases where multiple observations correspond to a single state. The proposed method was then used to estimate the displacement distribution of a bridge-abutment-embankment section that was known for its complex behavior during train passage. This behavior had not yet been fully clarified due to measurement difficulties. Compared with simply joining the displacements obtained from each camera, the results show that the proposed method produces a smooth displacement distribution at the camera boundaries, prevents phase shifts caused by time synchronization errors, and quantifies observation errors at each location.