SEISMIC MULTI-ATTRIBUTE FUSION USING FAST　INDEPENDENT COMPONENT ANALYSIS AND ITS　APPLICATION

摘要

Basic principles of independent component analysis (ICA) and fast independent component analysis (FastICA) algorithm are elaborated, and we propose an automatic fusion method of seismic multi-attribute based on FastICA. This method can calculate the transform kernel matrix rapidly using FastICA algorithm to achieve the feature fusion of several seismic attributes in the ICA domain. After that we map the synthesized attribute to the spatial domain to obtain the fusion result. Our method can remove the correlation hidden in high-order statistical characteristics between features. Finally, the application of 3D seismic data in northeastern Sichuan shows the effectiveness and rationality of the proposed method.