14-17 May 2018
New Orleans
US/Central timezone

Comparing Segmentation Methods for X-ray CT Images of Porous Media

17 May 2018, 12:15
New Orleans

New Orleans

Poster GS 2: Computational challenges in porous media simulation Poster 4


Mr Fei Zhao (CNPC Drilling Research Institution)


In order to understand the pore structures and fluid distribution in the oil reservoir rock which is a kind of porous media, X-ray computed tomography which can yield high-resolution, three-dimensional representations of pore space and fluid distribution within porous materials is applied to obtain data on structure, porosity, permeability, and other rock properties. Segmentation method of CT scan data is the crucial step for characterization and quantitative analysis of pore structures. Gray-scale density images obtained by CT scans of rock samples can be converted into binary images of grains and pores using segmentation techniques. The segmentation technique is applied to each 2D slice of a 3D gray-scale image to produce 2D binary sections. There are a vast number of segmentation methods in the world. This paper analyzes 10 different segmentation methods applicable to porous media for CT scan images of shales and oil sands. The porosity be derived by image applying the segmentation methods is also compared to the laboratory directly measured core porosity. The reason demonstrates that the application of different segmentation methods as well as associated operator biases yield vastly differing results. The quality of the results obtained using Rosenfeld’s method is not a reliable method. Otsu’s method gives the best results for shale samples and Kittler’s methods give fairly good results for oil sands samples. This illustrates the importance of the segmentation step for quantitative pore space analysis and fluid dynamics modeling.

Acceptance of Terms and Conditions Click here to agree

Primary author

Mr Fei Zhao (CNPC Drilling Research Institution)

Presentation Materials

There are no materials yet.