A sound understanding of the microstructure of dairy products is of great importance in order to predict and control their properties and final quality. The aim of this study was to develop an automated image analysis procedure to characterize the microstructure of different dairy systems. A high pressure freezing coupled with freeze-substitution (HPF-FS) protocol was applied prior to transmission electron microscopy (TEM) in order to minimize any modification of the microstructure of the dairy systems investigated. The developed image analysis procedure was first validated on synthetic images of suspensions, and then on two types of concentrated milk suspensions. Microstructural data relating to casein micelles in milk suspensions were taken from the literature. The established procedure was then applied to the two corresponding rennet-induced milk gels, prepared from the same milk concentrates used for suspensions preparation. The automated image analysis procedure allowed the reliable estimation of several characteristic microstructural parameters including area fraction, porosity, boundary length per unit area, particle aggregate size, inter-particle aggregate distance and tortuosity. The relative ease of estimating these microstructural parameters from the automated image analysis method could make it useful for routine measurements of milk gels. Moreover, the method enabled a useful discrimination between two different types of milk gels. This novel approach can contribute to a better understanding of the effects of processing on the structure-property relationships in dairy products, and may be applied to other food systems.