A New Image Registration Method Using Intensity Difference Data on Overlapped Image

On Image registration, it is intuitive to orientate images by matching corresponding pixels being considered idealistically identical on overlapping region. Based on this concept, the present paper proposes an image registration method that applies the information theorem to the corresponding intensity data. An entropy-based objective function is designed upon the histogram of the intensity differences. Intensity differences represent the differences of the corresponding pixels between the referenced and sensed images on the overlapping region. The sensed image is aligned to the referenced image by minimizing the entropy of the intensity differences through iteratively updating the parameters of the similarity transformation in the optimization process. The experimental results obtained by means of several test images illustrate the effectiveness of the proposed image registration method.