| Popis: |
Summary This paper describes a workflow for automatic processing of whole core images that enables macroscopic description of the rock layers constituting the studied geological section. The method makes it possible to detect core properties descriptors including, rock color characteristics, texture, layering inclination and thickness, and the shape and size of clastic inclusions. The method includes procedures for the detection of layering via convolution transformation of the image, clastic inclusions detection and characterization using watershed via segmentation procedures, processing color/intensity analysis, cluster analysis using normalized color/intensity data of a given image zone, and other image processing and analysis procedures, which enables obtaining any information for characterizing core samples. The parameters can be selected automatically or set manually. They depend on the particular task and target scale. Both normalizing daylight and UV images of core can be used. Clustering enables automatic detection and filtering of zones of photographs that are filled with organic spacers, heavily fractured material, or zones filled with organic matter. Processing results for terrigenous and carbonaceous whole core images are presented in the paper. The results are compared with petrophysical analysis of the rock material and profiling of the physical properties of the whole core materials. |
