{"title":"基于 \"自我对自我\"(Self2Self)和 \"丢弃\"(Dropout)的 CRP 采集的噪声抑制","authors":"Fei Li, Zhenbin Xia, Dawei Liu, Xiaokai Wang, Wenchao Chen, Juan Chen, Leiming Xu","doi":"arxiv-2408.02187","DOIUrl":null,"url":null,"abstract":"Noise suppression in seismic data processing is a crucial research focus for\nenhancing subsequent imaging and reservoir prediction. Deep learning has shown\npromise in computer vision and holds significant potential for seismic data\nprocessing. However, supervised learning, which relies on clean labels to train\nnetwork prediction models, faces challenges due to the unavailability of clean\nlabels for seismic exploration data. In contrast, self-supervised learning\nsubstitutes traditional supervised learning with surrogate tasks by different\nauxiliary means, exploiting internal input data information. Inspired by\nSelf2Self with Dropout, this paper presents a self-supervised learning-based\nnoise suppression method called Self-Supervised Deep Convolutional Networks\n(SSDCN), specifically designed for Common Reflection Point (CRP) gathers. We\nutilize pairs of Bernoulli-sampled instances of the input noisy image as\nsurrogate tasks to leverage its inherent structure. Furthermore, SSDCN\nincorporates geological knowledge through the normal moveout correction\ntechnique, which capitalizes on the approximately horizontal behavior and\nstrong self-similarity observed in useful signal events within CRP gathers. By\nexploiting the discrepancy in self-similarity between the useful signals and\nnoise in CRP gathers, SSDCN effectively extracts self-similarity features\nduring training iterations, prioritizing the extraction of useful signals to\nachieve noise suppression. Experimental results on synthetic and actual CRP\ngathers demonstrate that SSDCN achieves high-fidelity noise suppression.","PeriodicalId":501270,"journal":{"name":"arXiv - PHYS - Geophysics","volume":"7 1","pages":""},"PeriodicalIF":0.0000,"publicationDate":"2024-08-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Noise Suppression for CRP Gathers Based on Self2Self with Dropout\",\"authors\":\"Fei Li, Zhenbin Xia, Dawei Liu, Xiaokai Wang, Wenchao Chen, Juan Chen, Leiming Xu\",\"doi\":\"arxiv-2408.02187\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Noise suppression in seismic data processing is a crucial research focus for\\nenhancing subsequent imaging and reservoir prediction. Deep learning has shown\\npromise in computer vision and holds significant potential for seismic data\\nprocessing. However, supervised learning, which relies on clean labels to train\\nnetwork prediction models, faces challenges due to the unavailability of clean\\nlabels for seismic exploration data. In contrast, self-supervised learning\\nsubstitutes traditional supervised learning with surrogate tasks by different\\nauxiliary means, exploiting internal input data information. Inspired by\\nSelf2Self with Dropout, this paper presents a self-supervised learning-based\\nnoise suppression method called Self-Supervised Deep Convolutional Networks\\n(SSDCN), specifically designed for Common Reflection Point (CRP) gathers. We\\nutilize pairs of Bernoulli-sampled instances of the input noisy image as\\nsurrogate tasks to leverage its inherent structure. Furthermore, SSDCN\\nincorporates geological knowledge through the normal moveout correction\\ntechnique, which capitalizes on the approximately horizontal behavior and\\nstrong self-similarity observed in useful signal events within CRP gathers. By\\nexploiting the discrepancy in self-similarity between the useful signals and\\nnoise in CRP gathers, SSDCN effectively extracts self-similarity features\\nduring training iterations, prioritizing the extraction of useful signals to\\nachieve noise suppression. Experimental results on synthetic and actual CRP\\ngathers demonstrate that SSDCN achieves high-fidelity noise suppression.\",\"PeriodicalId\":501270,\"journal\":{\"name\":\"arXiv - PHYS - Geophysics\",\"volume\":\"7 1\",\"pages\":\"\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2024-08-05\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"arXiv - PHYS - Geophysics\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/arxiv-2408.02187\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"arXiv - PHYS - Geophysics","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/arxiv-2408.02187","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
摘要
地震数据处理中的噪声抑制是促进后续成像和储层预测的关键研究重点。深度学习在计算机视觉领域大有可为,在地震数据处理方面也具有巨大潜力。然而,依赖于干净标签来训练网络预测模型的监督学习面临着挑战,因为地震勘探数据无法获得干净标签。相比之下,自监督学习(self-supervised learning)通过不同的辅助手段,利用内部输入数据信息,以代用任务取代传统的监督学习。受 "自我对自我"(Self2Self with Dropout)的启发,本文提出了一种基于自我监督学习的噪声抑制方法,称为 "自我监督深度卷积网络"(Self-Supervised Deep Convolutional Networks,SSDCN),专门用于共反射点(CRP)采集。我们利用输入噪声图像的一对伯努利采样实例替代任务,以充分利用其固有结构。此外,SSDCN 还通过正常偏移校正技术纳入了地质知识,该技术利用了在 CRP 采集中有用信号事件中观察到的近似水平行为和较强的自相似性。通过利用 CRP 采集中有用信号与噪声之间的自相似性差异,SSDCN 在训练迭代过程中有效地提取了自相似性特征,优先提取有用信号以实现噪声抑制。在合成和实际 CRP 收集上的实验结果表明,SSDCN 实现了高保真噪声抑制。
Noise Suppression for CRP Gathers Based on Self2Self with Dropout
Noise suppression in seismic data processing is a crucial research focus for
enhancing subsequent imaging and reservoir prediction. Deep learning has shown
promise in computer vision and holds significant potential for seismic data
processing. However, supervised learning, which relies on clean labels to train
network prediction models, faces challenges due to the unavailability of clean
labels for seismic exploration data. In contrast, self-supervised learning
substitutes traditional supervised learning with surrogate tasks by different
auxiliary means, exploiting internal input data information. Inspired by
Self2Self with Dropout, this paper presents a self-supervised learning-based
noise suppression method called Self-Supervised Deep Convolutional Networks
(SSDCN), specifically designed for Common Reflection Point (CRP) gathers. We
utilize pairs of Bernoulli-sampled instances of the input noisy image as
surrogate tasks to leverage its inherent structure. Furthermore, SSDCN
incorporates geological knowledge through the normal moveout correction
technique, which capitalizes on the approximately horizontal behavior and
strong self-similarity observed in useful signal events within CRP gathers. By
exploiting the discrepancy in self-similarity between the useful signals and
noise in CRP gathers, SSDCN effectively extracts self-similarity features
during training iterations, prioritizing the extraction of useful signals to
achieve noise suppression. Experimental results on synthetic and actual CRP
gathers demonstrate that SSDCN achieves high-fidelity noise suppression.