A two-step mechanism for automated 3D horizon picking
Feng Qian*, Chang Luo, Zhaojie Su, Xingmiao Yao, Guangmin Hu, Xiaoer Cheng · 2014
Summary The spatial distribution characteristics of seismic extrema in 3D seismic data are rendered that extrema are high density in the same horizon, while density of different horizons is low in time direction. Based on this understanding, we propose a novel automatic tracking algorithm that mainly contains two steps: (1) form horizon fragments using density-based spatial clustering; (2) merge fragments into complete horizon surface by finite Gaussian mixture model based clustering algorithm.