Abstract

Seismic attributes play the most effective role in the interpretation to predict geological features from seismic images. However, using a single attribute for the purpose may reduce the prediction quality. Therefore, multiple attributes analysis becomes significant and has the potential to decipher the finer details. In recent days the Principal Component Analysis (PCA) is used to generate the multi-attribute section. As the PCA is a linear method, it can only map the data to linear Principal Components. Generally, seismic attributes exhibit nonlinear relationships and the seismic data contains nonlinearity in complex geological structures. So, the linear PCA may not map the attributes well. So here we propose an Auto associative neural network-based Non-Linear Principal component analysis method. In this, the principal components are generalized from straight lines to curves. It has the ability to map the nonlinearities in the attributes, so it can map the geological complexities well. We applied our method in the Andaman offshore seismic data and the resulting multi-attribute section enhanced both structural and stratigraphic features