A More Efficient and Robust Scheme for the Implementation of Eigenstructure-Based Coherence Computation

When seismic data is of low signal-to-noise ratio (SNR), more traces shall be involved to improve robustness against noise, and thus a larger analysis window shall be used. However, this will lead to increased time cost and blurred coherence images. In this paper, we propose a modified eigenstructure-based coherence (C3m) algorithm based on inverse distance weighting method and dimensionality reduction of the covariance matrix for eigendecomposition in the case of a large analysis window size exceeding 3 by 3 (9 traces). In the case of large analysis windows, the time cost by C3m is at least 30% less than C3 via synthetic data and field data studies, and the coherence results calculated with C3m are clearer than those calculated with C3. The new method may be an efficient alternative to C3 for coherence computation.