Detection of reservoir fractures in imaging logs using directional filtering

Every rock, which has pores connected together, can be considered as a reservoir rock. Existing fractures in the reservoir rock are the most effective factors, which increase the production and harvesting of the well. Detection of reservoir rock fractures is a very complicated process, due to factors such as poor contrast, fracture variable thickness, noise, discontinuities in the curve of a fracture, and the existence of some phenomenon like fracture bedding. In the well-log images, some sinusoidal curves are fractures. The directions of fractures are shown as some special curves. According to this concept, in this paper, fractures are extracted based on an optimized algorithm using directional filters. The first order derivative of the imaging log is calculated for each pixel. Then, the direction filters are designed based on a different direction of each pixel, and applied to the original imaging log. Finally, using the Hough transform, some parameters of each fracture are obtained, and are displayed on a real imaging log.