Abstract: In this study, to reduce the impact of non-measureable noise on the measurement accuracy and to improve the monitoring accuracy of an engineering structure′s dynamic deformation in digital photography, the standard Kalman filter, variance-compensated Kalman filter, maximum posterior Kalman filter, and variance component Kalman filter are used to deal with the bridge elastic large deformation data; the Kalman filtering adaptability in dealing with dynamic deformation data is studied; and the advantages of employing variance component Kalman filter in data processing are quantified. It is found that the noise processing of variance component Kalman filter is stable and the error is small. This method is suitable for the processing of dynamic elastic small deformation noise in digital photography and exhibits good results in the processing of dynamic elastic large deformation noise in digital photography. Laboratory experiments on similar materials show that after data processing via variance component Kalman filter, the measurement error is less than 0.5 mm, which meets the accuracy requirement of deformation monitoring.

Key words: digital photography, deformation monitoring; Kalman filter