X-ray computer tomography scanning (X-ray CT) has been adopted for investigating rock's microstructure, which could characterize sample's inner pore structure non-destructively. Image segmentation is the first step towards pore space identification. Segmentation of a porous media means conversion of gray scale CT volumes into different regions that permits quantitative characterization. Macroscopic properties of the segmented image can vary greatly with different segmentation methods. In general, CT images of rock and soil samples are required to convert into binary images of particles and pores. X-ray computer tomography technique could also be used for visualizing the specific bitumen-pore-grain structure of oil sands. Segmentation of oil sands refers to the process of partitioning the image into three regions of void, grains and bitumen. At present, the applicability investigation of various segmentation methods for oil sands seems to lag.
In this study, four methods for identifying bitumen and pore regions from 2D oil sands Micro CT scan images are compared. Image information entropy is used as a weight to calculate the average porosity of the sample. Image-derived porosity obtained according to the segmentation are given in this paper. Several technical issues of image preprocessing of CT images of oil sands are discussed, and the porosity and bitumen content derived from image applied the segmentation methods are compared to the laboratory measurement of oil sands cores. The research presented is the foundation of quantitative pore space analysis and hydraulic modeling of oil sands based on digital cores.
Oil sands is a nature mixture of sand, clay, water, and extra heavy crude oil, which is called bitumen. The most effective oil sand bitumen thermal recovery technology is the steam assisted gravity drainage (SAGD) process. Bitumen is immobile at reservoir temperatures and becomes less viscous or even begin to flow with the increase of temperature. To understand the characterization of pore structures and pore networks of oil sands is regard as a significant approach to evaluate the mechanical and hydraulic properties.