Image-based microscale rock typing and its application

Abstract

Abstract Rock typing plays a crucial role in describing the heterogeneity of the reservoir. Most of the conventional rock typing methods are implemented to classify the target reservoir into various rock types based on various petrophysical properties (e.g., porosity and permeability), but fail to provide more critical information that significantly affects the final performance of the reservoir characterization including: (1) the porosity and permeability contribution of each rock type and (2) the geological genesis of each rock type. Along with the universal application of various imaging devices, the image-based microscale rock typing (IMRT) can be directly conducted based on the observed pore structures which fundamentally determine the rock types. The IMRT belongs to the computer vision field which can be divided into pattern recognition-related rock typing (PRRT) and texture segmentation-related rock typing (TSRT). The PRRT is mainly used to identify the category (e.g., lithofacies, reservoir zone, or Dunham textures) of a given rock sample. The TSRT aims to classify a single image into several areas where each area denotes a relatively homogeneous porous structure. In this paper, the popular IMRT methods and their applications are reviewed thoroughly. Many successful applications proved that IMRT is an effective way to quantitatively estimate the porosity and permeability contributions of each rock type in a heterogeneous rock sample with the help of numerical flow simulation. Besides, the IMRT results also can be used to reveal the geological genesis of each rock type when its texture is determined by a special geological process.

Keywords

• Rock typing
• Digital rock physics
• Texture segmentation
• Pore-scale reservoir characterization