人工智能地震分类研究进展综述  

Overview of progress in research on artificial intelligence for earthquake classification

作　　者：陆天然 段梦乔 李子怡 周连庆[1] LU TianRan;DUAN MengQiao;LI ZiYi;ZHOU LianQing(Institute of Earthquke Forecasting,China Eaerthquake Administration,Beijing 100036,China)

机构地区：[1]中国地震局地震预测研究所,北京100036

出　　处：《地球物理学进展》2025年第1期25-47,共23页Progress in Geophysics

基　　金：国家重点研发计划(2021YFC3000704);国家自然科学基金项目(42174066);中国地震局地震预测研究所基本科研业务费专项(CEAIEF20230210)联合资助.

摘　　要：各类地震的正确分类对区域地震危险性评估,降低天然地震或人工地震灾害等方面的工作具有重要意义.多年来,许多科技人员对此进行了大量研究.本文较为系统地总结了国内外目前对天然地震和人工地震各类型事件分类特征及难点的主要认识,以及人工智能在地震分类研究中的应用现状.结果表明:(1)爆破地震与构造地震的分类已取得较快进展;(2)火山地震与滑坡事件的识别分类普遍存在数据集样本数量过少和数据不平衡的问题;(3)诱发地震的识别仍是分类研究中较为困难和存在争议的课题;(4)人工智能方法以高准确率、高效率和未来自动化的可能性成为目前主流的地震分类方法.结合人工智能技术在地震分类中的应用现状,本文讨论并提出人工智能地震分类的趋势与相应建议.The correct classification of earthquake events is of great significance to regional seismic hazard assessment and the reduction of natural or artificial earthquake disasters.Over the years,many scientific and technological personnel have conducted a large amount of research on this topic.This article systematically summarizes the current mainstream understanding of the classification characteristics and difficulties of various types of natural and artificial earthquake events in China and abroad,as well as the application status of artificial intelligence in earthquake classification research.The results show that:(1)the classification of blasting earthquakes and tectonic earthquakes has made relatively rapid progress;(2)the identification and classification of volcanic earthquakes and landslide events generally face the problems of insufficient sample size and data imbalance in the dataset;(3)the recognition of induced earthquakes remains a difficult and controversial subject in classification research;(4)artificial intelligence methods,with their high accuracy,efficiency,and potential for future automation,have become the mainstream method for earthquake classification at present.Based on the current application status of artificial intelligence technology in earthquake classification,this paper discusses and proposes corresponding suggestions and development trends.

关 键 词：天然地震 爆破 诱发地震 火山地震 地震分类 人工智能 

分 类 号：P315[天文地球—地震学]