Image mosaicing can solve the problem that only very limited field of view is available when highly magnifying lens is used in microscope imaging. This paper puts forward an equation of the system's influence to the grayscale distribution of the medical microscopic images, based on the analysis of the influence. Then, by the system equation and the accurate registration equation, the least-squares optimization error function can be formulated. By these equations, a set of transcendental equations is formed. We can estimate the system's influence on the gray-level of the images by singular-value decomposition (SVD), based on the equations set, and compensating for it. This permits image mosaicing seamlessly, without lowering the speed.