摘要
对深度学习的方法原理及主流的前馈神经网络、卷积神经网络、循环神经网络、Transformer网络、自编码器、生成对抗网络以及深度强化学习网络等进行了介绍,讨论了不同网络的适用领域。从震相拾取、震相关联、地震定位与事件检测,地震信号和地震事件的分类,地震预测预报等方面对近年来深度学习方法的应用技术进行了提炼总结,综述了深度学习方法的应用进展,讨论了当前常见深度学习方法在地震监测预报领域中的主要应用方式、优势特点及解决的主要问题。总结了现阶段深度学习方法在地震监测预报领域中存在的应用局限性以及后续发展方向。
This article provides an overview of deep learning methods and their application in earthquake monitoring and prediction.It introduces mainstream methods such as feedforward neural networks,convolutional neural networks,recurrent neural networks,transformer networks,autoencoders,generative adversarial networks,and deep reinforcement learning networks.The article summarizes their application in phase picking,phase correlation,event detection,earthquake location,signal and event classification,and earthquake prediction.It also discusses the progress,advantages,challenges,and future directions of deep learning in earthquake monitoring and prediction.This summary serves as a valuable reference for applying deep learning in earthquake monitoring and prediction.
作者
贾漯昭
孟令媛
闫睿
JIA Luozhao;MENG Lingyuan;YAN Rui(Henan Earthquake Agency,Zhengzhou 450018,Henan,China;China Earthquake Networks Center,Beijing 100045,China)
出处
《地震研究》
CSCD
北大核心
2024年第3期336-349,共14页
Journal of Seismological Research
基金
国家重点研发计划(2021YFC3000705)
中国地震局震情跟踪定向工作任务(2023010111).
关键词
深度学习
监测预报
地震检测
地震定位
地震预测
deep learning
seismic monitoring and prediction
earthquake detection
earthquake location
seismic forecasting
