Automatic Extraction of Fractures in Digital Geological Outcrop of Carbonate Rocks and Its Application

Abstract

Fracture identification and characterization is an important part of fractured reservoir research. To realize the automatic extraction of digital geological outcrop fractures, a fracture extraction method based on multi-scale Beamlet transform is designed. Firstly, the image is enhanced by the adaptive enhancement algorithm to eliminate the influence of noise and uneven illumination. Then the fracture binary image obtained by the Otsu threshold method is used to extract the fracture by Beamlet transform. Finally, discrete line segment connection algorithm is designed to complete the fracture connection. Experiments show that, compared with Canny edge detection algorithm and Hough transform algorithm, the proposed method has better anti-noise ability, and the extracted fracture linear features are more accurate and complete. By adopting adaptive enhancement processing and discrete line segment connection algorithm, the problems of uneven illumination and disconnection of line segments are reduced, and the advantages of Beamlet algorithm in linear feature extraction are fully utilized. Through the analysis of the application of this method on digital outcrop of the Member-2 of Sinian Dengying Formation in Ebian Xianfeng area, it is shown that this method has a good practical application potential in large-scale fracture extraction and complements the study of microscopic fracture identification. It has a good application prospect for improving the interpretation efficiency of digital geological outcrop fractures and has reference value for other linear feature extraction applications.