Abstract: Aircraft icing is formed dynamically and appears with porous structure in microcosmic view. This structure directly influences the physical properties of ice, and varies with different icing conditions. In order to know this influence law better, the paper focuses on the difficult about question how micro features changed quantitatively at different temperatures. At the beginning, an information extraction approach is presented based on the image segmentation method. Secondly, some assumptions are given following the quantitative results of ice in microcosmic view, and then the 3D mathematical model is built based on them. In order to measure distribution characteristics of air pores in the model quantitatively, a computational density is proposed. Finally, take micro-images of ice formed in icing wind tunnel as objects, the feasibility and rationality of the proposed information extraction method and mathematical model are demonstrated. Pores distribution and computational density of ice at different temperatures are also discussed, and the results show highly consistent with the macroscopic viewpoints. Through this paper, it provides quantitative basis for deep understanding of physical phenomenon and properties, and a new way to ice study in microscopic view can be given.