Fracture and vug characterization and carbonate rock type automatic classification using X-ray CT images

This paper presents a method of fracture and vug characterization and automatic classification of rock type in a naturally fractured vuggy carbonate reservoir using X-ray computed tomography (CT). First, using image processing technologies such as image segmentation, morphological processing and skeleton extraction, the fractures and vugs in the CT images of core plugs were characterized, and then, the characteristic parameters of the fractures and vugs were calculated. Then, a support vector machine (SVM) was utilized for automatic classification of the fractures and vugs. The shape and the eccentricity of fractures and vugs as the SVM training eigenvectors can effectively distinguish fractures from vugs with high classification confidence. Finally, with the characteristic parameters of fractures and vugs calculated from the CT images of core plugs, a new defined function of the carbonate rock index (CRI) was proposed, which can quantitatively classify the carbonate rock into three different types: matrix, fractured and vuggy. After analysis of more than 200 core plugs from the Y reservoir in the Middle East, the CRI values for matrix, vuggy and fractured carbonate rock were given and are 0 < CRI < 1, 1 <= CRI <= 10 and CRI > 10, respectively. A high CRI value indicates a high fracture density, and a small CRI value indicates that the carbonate rock is mainly matrix without fractures and vugs. For the vuggy carbonate core plug, the CRI value is in between those of fractured core plugs and matrix core plugs. Therefore, the value of CRI can quantitatively evaluate the heterogeneity of carbonate rock due to fractures and vugs. The field application of this method demonstrated that this new method can effectively characterize fractures and vugs and quantitatively classify the rock type of the naturally fractured vuggy carbonate reservoir.