Estimating Seismic Dip Fields with Steerable Filters

We have developed a novel dip estimation method using a family of smooth oriented filters known as steerable filters (Freeman et al., 1991). 3D dip attribute cubes are calculated without explicitly picking any horizons. Steerable filters based on nth Gaussian derivative kernels have demonstrated their abilities to process noisy data in a structure-oriented manner. We demonstrate this steerable filter processing solution allows user-controlled trade-off on the scale, smoothness and the accuracy of the estimation for noisy and discontinuous data. We believe our approach supports practical tools for 3D dip estimation, since it is structure-oriented and convenient to parallelize on accelerated computational platforms.