A mosaicing algorithm for series of microscope images based on minimum cost spanning tree

Available methods for a series of microscope images mosaicing can not meet the requirement of large scale microscope images mosaicing with high quality requirements. A mosaicing algorithm for a series of microscope images based on minimum cost spanning tree of mosaicing graph is proposed and implemented in 3 steps: a) the local registrations based on blocking-spatial clustering, b) the evaluation of the results of images registrations to assign weights of edges of the mosaicing graph, and c) the global images mosaicing approach based on minimum cost spanning tree to calculate optimal global position for every image in order to construct panorama in high quality. The algorithm can be implemented easily and features high accuracy in the experiments.