Advances in Vug Quantification: Leveraging Adaptive Thresholding, Gaussian Weighting, and Laplacian Contrast Analysis in Borehole Images

Summary Image logs are crucial in capturing high-quality geological information about subsurface formations. Among the array of geological features identifiable through borehole images (BHIs), vugs, together with fractures, stand out as one of the critical features for assessing reservoir quality in carbonate rocks. This study presents an innovative methodology for the detection and statistical analysis of vugs from image logs. Using a series of advanced computer vision and image processing techniques, this methodology uses mode extraction, local variation enhancement, Gaussian-modulated adaptive thresholding, contour extraction, and refinement based on geometric criteria. Further complemented by Laplacian-based contrast and mean-value-based filtering, the proposed methodology was able to successfully perform the task of individual vug identification, quantification, and statistical analysis. Detailed metrics, such as total count, total area, mean, and standard deviation of vug areas every 0.1-m interval, were introduced. A detailed statistical analysis of the interpreted vugs provides insights into their area, circularity, and resistivity distributions, along with porosity, circularity, and azimuth spectrum, offering a deeper understanding of the reservoir characteristics. This method was compared with vugs extracted by the expert from the image log interpretation software, WellCAD, showing comparable vug extraction and, at times, identifying vugs missed by the expert. Additionally, the method enables extra statistical analysis and individual vug quantification unavailable in WellCAD and similar tools. Applied to a case study from an Oman carbonate reservoir, the method showcases improvement in processing speed, alongside improvement in accuracy, individual vug qualification, and the granularity of vugs characterization, when compared with conventional manual approaches. The method’s effectiveness was also tested on different well orientations, including both vertical and horizontal wells, as well as its adaptability to varying image log types, specifically formation microimager (FMI) and compact microimager (CMI) logs. The findings suggest that this approach can substantially aid in reservoir evaluation, offering a rapid, detailed, and accurate tool for geological studies.