Seismic profile denoising based on common-reflection-point gathers using convolution neural networks

Abstract With the development of seismic surveys and decline shallow petroleum resources, high resolution signal-to-noise ratio have become more important in processing. To improve quality data, stationary-phase migration based on dip-angle gathers can be used to separate reflected waves noise. However, this method is very computationally intensive heavily dependent expert experience. Neural networks currently powerful adaptive capabilities great potential replace artificial Certain applications convolution neural (CNNs) stack profiles lead a loss amplitude information. Therefore, we developed CNNs for noise reduction common-reflection-point (CRP) gathers. We CRP as labels conventional prestack time inputs. In addition, analyzed properties demonstrated network optimization process results. The results showed that our methods achieve fast reliable denoising produce high-quality contain true Furthermore, predicted further processing steps, such normal moveout correction variation with offset.