Advanced 3-D imaging techniques such as X-ray Computed Tomography (CT) in association with mathematical morphology based analysis can be employed to quantify the geometrical and topological features of pore networks. The first and critical step in image analysis is binarization to separate pores from the surrounding soil matrix. Results of all subsequent two- and three-dimensional analyses is strongly dependent on the accuracy of the binarization step, hence the choice of thresholding technique is of crucial importance. In this study we tested several global and local thresholding techniques for binarization of CT images of geological materials. Geometrical and topological pore network features such as porosity, specific surface area, pore size distribution, tortuosity, and connectivity were determined based on Mathematical morphology operations.