Shale is a complex porous medium with low porosity, super-low permeability, tight and heterogeneity pore structure. The distribution of each component within shale can be visually revealed by the three-dimensional digital cores. Moreover, microscopic structural characterization and seepage simulation experiments of shale can also be carried out repeatedly and nondestructively using digital core. Therefore, the digital core is a more effective and convenient method than the conventional physics experiment methods for shale. The digital core constructed by the physical construction method is more accurate, however, which can only reveal the physical properties of a particular rock sample, and is generally effected by the contradiction between the field of view and resolution. Thus, it is necessary to construct the simple digital core models with a certain universality, which is constrained by some parameters, and can represent the characteristics of a kind of lithology or lithofacies. In the meantime, the current image segmentation methods are difficult to synthesize the multi-sample characteristics, which represent the characteristics of a kind of lithology or lithofacies. In this paper, we propose a method about threshold segmentation. It segmentalizes CT images based on the cumulative distribution function of image pixels, combined with the percentage of each component groups and the gray level distribution sequence. The volume relative content and gray level distribution sequence of mineral compositions and porosity of different lithofacies shale in Dongying depression are obtained by means of scanning electron microscopy (SEM) and energy-dispersive X-ray spectroscopy (EDX). And then the digital core models of different lithofacies shale, including rich felsic, felsic, rich calcareous, calcareous and rich argillaceous lithofacies shale, are constructed based on the X-ray computed tomography (CT) imaging. The results show that: (1) According to the grey value, the six component groups of shale reservoirs in Dongying depression from low to high,are followed by pore-fractures, organic matter, clay minerals, felsic minerals, calcium minerals and iron minerals, respectively;(2) Cumulative distribution function of pixels in CT gray image is represented cumulative distribution function of volume relative content of different lithofacies sample component groups. Based on volume relative content and gray level distribution sequence of different lithofacies shale sample components, threshold of each component group can be obtained in turn. This method is more suitable for threshold segmentation of specific lithofacies, while other segmentation methods cannot achieve; (3) The simple standard digital core models can be constructed by a superposition method. Firstly, the single-component digital cores are constructed based on CT images and thresholds of various petrographic component groups. And then these digital cores are superimposed together and the complete simple digital core is reconstructed, which is constrained by the parameters, such as mineral components and porosity. The simple standard digital core models, which have a certain universality, can represent the characteristics of a kind of lithology or lithofacies.
