Geologic-target amplitude and edge preserving oriented random noise elimination
Yao Zhen-an, Chengyu Sun, Jie Tang · SPG/SEG 2016 International Geophysical Conference, Beijing, China, 20-22 April 2016 · 2016
Summary Curvelet transform thresholding method is effective in random noise elimination based on threshold selection, and it results in excessive smoothing and surrounding effect at the same time. Anisotropic diffusion filtering can preserve the border structures of seismic texture images efficiently, which always applied in image enhancing. Based on the analyis of Curvelet thresholds, an new effective method was presented in this paper which combined the advantages of both curvelet transform and anisotropic diffusion. The new method can remove random noise and preserve edge structures of seismic events at the same time, which was certified by the test of theoretical and actual seismic data. Advantages on improving signal to noise ratio and edge preserving were analyzed through six kinds of quantitative evaluation indexs.