Abstract
▼

Multiple reflections(multiples) are the main causes of misinterpreting of primary reflections in the marine seismic reflection data, both in shot gathers and stack sections. Thus, they need to be suppressed throughout data processing. Conventional methods such as prediction-error deconvolution and R...

Multiple reflections(multiples) are the main causes of misinterpreting of primary reflections in the marine seismic reflection data, both in shot gathers and stack sections. Thus, they need to be suppressed throughout data processing. Conventional methods such as prediction-error deconvolution and Radon filtering have been used for attenuating the multiples. However, the vast majority of processing workflows require time-consuming steps to testing and selecting the processing parameters, along with skilled data processing techniques. To attenuate the multiples in seabed seismic reflection data, we utilize deep learning system with AutoEncoder model and the U-net model. The improved S/N with rejection of the multiples is investigated in common shot gathers, CMP gathers, NMO-corrected CMP gathers, and stack sections. Compared to AutoEncoder model, U-net model is found to be more effective with regard to image detection and attenuation of multiples in seabed seismic reflection data

주제어

#seabed seismic reflection data multiples machine learning deep learning Auto-encoder model U-net model;

참고문헌 (0)

1. 이 논문의 참고문헌 없음

이 논문을 인용한 문헌 (0)

1. 이 논문을 인용한 문헌 없음