A Registration and Change Detection Algorithm Based on Pre-and Post-Strike Image

Zihe Liu, Yuhui Wang

Abstract: Image-based damage assessment requires processing and analysis of two images taken at the same location and at different times. For the registration of this type of image, the traditional method based on gray information is not applicable due to the change of the image itself. Image registration based on Fourier-Mellin Transform is a global and phase correlation method which is carried on Cartesian coordinate and Log-polar coordinate. This method extracts the extreme value of the impulse function of the image phase in the frequency domain, and obtains the transformation parameters of the image. In combination with the SIFT registration method, the effect of image registration with a large rotation angle is improved. Experiments show that the algorithm is effective and feasible. Based on the registration, the difference method is used to extract the image change area. The adaptive threshold method is used to preliminarily determine the interested targets, which provides a basis for the establishment of the damage model in the next step.