Improving the eigen-structure seismic attribute in identifying seismic discontinuities using eigenvectors

Faults and various types of discontinuities play a significant role in the exploration, extraction, and exploitation of hydrocarbon resources. One of the important and widely used tools for identifying these geological features is seismic data. Although faults create discontinuities in seismic data, making their identification possible, due to various reasons such as the detrimental effects of noise and the time-consuming nature of manual identification, seismic attributes are now commonly used for fault detection and exploration in seismic data. Various types of attributes have been introduced for identifying seismic discontinuities, among which coherence attribute is one of the most widely used and common. In this research, a new coherence attribute based on eigenvectors and eigenvalues of the covariance matrix is examined and analyzed. This method, in comparison to the conventional structure-based coherence attribute utilizes both eigenvalues and eigenvectors for coherence analysis. The use of eigenvectors increases the method's sensitivity to changes in shape and polarization, whereas the conventional structure-based method, which relies solely on eigenvalues, is sensitive to energy changes. The results from evaluating the proposed attribute on synthetic and real data indicate that this method offers higher resolution in identifying discontinuities, particularly those that only cause waveform changes and polarization shifts, compared to semblance-based and conventional structure-based methods. Additionally, the analysis of results shows that the improved structure-based coherence attribute exhibits acceptable stability in the presence of random noise. These characteristics make the proposed method a more effective tool for identifying faults and discontinuities in seismic data.