Automatic salt dome detection in seismic data by combination of attribute analysis on CRS images and IGU map delineation

Abstract Accurate detection of geo-structures in automatic geological interpretation of seismic data is a challenging task especially in complex media. Solution to this problem goes through integration of advanced seismic pattern recognition methods, image analysis and image classification strategies, applied on seismic image obtained by appropriate seismic imaging method. Accuracy of these approaches mostly depends on the quality of seismic attributes used for subsequent image analysis, seismic data classification and seismic pattern recognition. In this study, we initially analyze the effect of using the common reflection surface (CRS) method on obtaining appropriate data for attribute analysis, in comparison with the conventional NMO/DMO/Stack imaging method. Subsequently, we introduce an automatic salt dome detection method from seismic image based on the intrinsic geological unit (IGU) concept used in sequential mineral exploration. The IGU consists of areas with indexes or critical genetic factors (CGF) related to the target of exploration which is defined itself by several critical recognition criteria (CRC). The CRCs defined as the value of attributes and seismic patterns selected by the interpreter for salt detection. These CRCs were extracted from a large informative database build only for salt dome detection from various seismic data and different geological setting. The CRCs were weighted by the interpreter according to the percentage of success in using selected CRCs for salt identification. Afterwards, probability value of pixels to be considered as salt was calculated by defining a linear relationship between CRCs. Then, matrix of characteristics was defined for CRCs followed by a matrix that shows CRCs score. The final score for each pixel will define the area of the IGU, or salt dome in seismic image, if they were above a predefined threshold. The presented strategy was applied on a field data example from Kazakhstan containing a huge salt dome. The proposed strategy could detect the salt boundary in a fully automatic manner with some befits, such as defining internal reflections and less human interaction. Application of the proposed method on the CRS result showed that it could be considered as an alternative to the other present automatic salt dome detection and classification approaches. However, the presented method still requires of using advanced method in scoring and building matrix of characteristics, especially for salt area detection in seismic image.