Automatic Detection of the Degree of Compaction in Reservoir-Rocks Based on Visual Knowledge

  • You are here
  • >
  • Home
  • >
  • Publications
  • >
  • Automatic Detection of the Degree of Com...

AAPG Annual Convention and Exhibition 2009

Paper presented at the 2009 AAPG Annual Convention & Exhibition. Denver, CO.

A low-cost method is proposed for evaluating the degree of compaction in reservoir rocks by using automatic inference methods on optical photomicrographs. In order to reproduce the visual interpretation performed during petrographic analysis, a hybridmethod was developed combining image processing algorithms with knowledge representation and reasoning models. The method proposed was inspired on visual attention, the mechanism used by the human brain for dealing with visual information. This mechanism allows the brain to filter the huge amount of information that comesthrough the eyes, selecting the relevant elements to be further analysed by the highly abstract level of reasoning. The process involves the decomposition of scenes, and the competition among their different aspects in order to isolate and select the relevantareas. In other words, the eyes of petrographers initially examine a thin-section by capturing and isolating the grains borders (outlines), and then focus on the grains. The outlines are essential to separate each grain from other grains and their interstices, because petrographic analysis is performed in the two-dimensional universe of thinsections. The knowledge at this level is modelled in terms of Sections (grains), Outlines (borders of grains), and Interstices, which may be Pores (empty) or NonPores (e.g. cement, matrix). The shapes of the outlines (mainly concave or convex), complemented by the detection of the impregnation blue resin, indicates if they contain Pores, NonPores or Sections. The types of contacts between grains are then used to define the degree of compaction of the rocks. The system provides a preliminary identification of the objects that can be interactively refined by the user when the grain outlines are unclear in the images. The evaluation of compaction degree provided by this method is far more sensitive and precise than those based on the intergranular volume or number of intergranular contacts. This formalized interpretation method shows better results for the complex tasks of reservoir quality characterization and prediction.