模型反演和深度学习反演联合的地震波阻抗优化反演
Seismic impedance optimization inversion combining model inversion with deep learning inversion
黄闻露 1阎建国 1任立龙 1谢锐1
作者信息
- 1. 成都理工大学地球物理学院,四川成都 610059
- 折叠
摘要
基于"数据驱动和模型驱动"相结合的思想,通过模型反演结果扩展标签训练集,并在深度学习算法中加入模型反演目标函数,对损失函数进行重构,提出了一种模型反演和深度学习反演联合的地震波阻抗优化反演.采用RNN网络结构实现了一种"伪标签"下的半监督深度学习网络反演,并用网络反演结果作为初始模型参与模型反演,最终优化反演由网络反演和模型反演不断迭代优化完成.通过合成Marmousi模型和实际资料,验证了所提出的方法具有较高的反演精度和实用性.
Abstract
Based on the combination ofdata-and model-driven approaches,this study expanded the labels of the training set through model inversion results,and added the model inversion objective function to the deep learning algorithm.By constructing a new loss function,this study proposed a seismic impedance optimization inversion method combining model inversion with deep learning inver-sion.The semi-supervised deep learning network inversion under a pseudo-label was achieved using the RNN network structure.The network inversion results were used as the initial model to participate in the model inversion.The final optimization inversion was com-pleted by continuous iterative optimization of both network and model inversion.The method proposed in this study proves to possess high inversion accuracy and practicability,as demonstrated by the synthesis of the Marmousi model and the actual data.
关键词
数据驱动/模型驱动/伪标签/半监督/波阻抗反演Key words
data-driven/model-driven/pseudo-label/semi-supervision/wave impedance inversion引用本文复制引用
基金项目
中国石油项目(RPED-2020-JS-121)
出版年
2024
国家科技期刊平台
NSTL
万方数据