Faults and fractures detection in 2D seismic data based on principal component analysis

Various approached have been introduced to extract as much as information form seismic image for any specific reservoir or geological study. Modeling of faults and fractures are among the most attracted objects for interpretation in geological study on seismic images that several strategies have been presented for this specific purpose. In this study, we have presented a modified approach of application concept of the principle components analysis to enhance faults and fractures from low quality seismic image. In the first step, relevant attributes considering imaging faults and fractures were have drawn based on vast study on previous successful applications of different attributes. Subsequently, major informative components of each attribute were defined by performing principle component analysis. Since random noise in seismic image exhibits no correlation in seismic data, true reflectors and diffraction events show high coherency value thus these objects would be separated into different orthogonal components in principle component analysis. It will make it easy to remove irrelevant information considering faults and fractures from seismic image and thus will make a higher quality image by combining attribute sections in principle component analysis. Afterwards, selected components were stacked to enhance the fault position in final image. However, since that are other geological objects that might show correlation in other orthogonal components, so there should be refinement step on the final image to stack only the favorable information. This approach was performed on a field land data example form north east of Iran. Result of application the proposed strategy shows that the method is capable to image faults compared to the conventional image analysis for fault detection. The method was also capable to image accurate position of the body of mud volcanoes exited in the image that could not be easily tracked by conventional seismic image analysis.