欢迎访问《地震》,

地震 ›› 2023, Vol. 43 ›› Issue (2): 166-188.doi: 10.12196/j.issn.1000-3274.2023.02.014

• • 上一篇    

深度学习在InSAR数据处理与地壳形变观测中的应用研究进展

张敬业1, 孙珂1,2, 张国宏2   

  1. 1.中国地震局地震预测研究所, 北京 100036;
    2.地震动力学国家重点实验室, 中国地震局地质研究所, 北京 100029
  • 收稿日期:2022-12-23 修回日期:2023-03-21 发布日期:2023-07-05
  • 通讯作者: 孙珂, 研究员。 E-mail: sunke@cea-ies.ac.cn
  • 作者简介:张敬业(1997-), 男, 山东临沂人, 在读硕士研究生, 主要从事人工智能地震预测研究。
  • 基金资助:
    国家重点研发计划项目(2019YFC1509202); 国家自然科学基金联合基金项目(U2039202)

Research Progress in Data Processing and Application of Deep Learning in InSAR Crustal Deformation Observation

ZHANG Jing-ye1, SUN Ke1,2, ZHANG Guo-hong2   

  1. 1. Institute of Earthquake Forecasting, China Earthquake Administration, Beijing 100036, China;
    2. State Key Laboratory of Earthquake Dynamnics, lnstitute of Geology, China Earthquake Administration, Beijing 100029, China
  • Received:2022-12-23 Revised:2023-03-21 Published:2023-07-05

摘要: InSAR(Interferometric Synthetic Aperture Radar, InSAR)技术凭借其高精度、 大范围、 全天候监测的优势, 在地表高程、 形变等信息的获取、 反演等应用中得到了广泛的认可, 并逐渐发展为地壳形变观测领域里不可或缺的技术手段, 但利用InSAR技术进行地壳形变观测离不开海量数据的支持, 这势必会给信息的收集和解读带来新的挑战。 近些年, 机器学习快速发展并在遥感图像处理方面取得了令人鼓舞的成绩, 将深度学习方法与InSAR技术相结合的尝试应运而生, 深度学习突出的数据挖掘能力和对目标任务的分类、 预测能力将会为InSAR数据处理和地壳形变观测中的应用提供新技术手段。 本文介绍了深度学习在InSAR数据处理与地壳形变观测中的应用研究主要进展, 并对其应用前景进行了讨论和展望。

关键词: 深度学习, InSAR, 地壳形变, 研究进展

Abstract: Interferometric synthetic aperture radar (InSAR) technology, with its advantages of high precision, large range and all-weather monitoring, has been widely recognized in the acquisition and inversion of surface elevation and deformation information and other applications, and has gradually developed into an indispensable technical means in the field of crustal deformation observation. However, the use of InSAR technology for crustal deformation observation cannot be achieved without the support of massive data. This is bound to create new challenges in the collection and interpretation of information. In recent years, the rapid development of machine learning has made encouraging achievements in remote sensing image processing. The attempt to combine the deep learning method with InSAR technology comes into being. The outstanding data mining ability of deep learning and the classification and prediction ability of target tasks will provide a new technical means for the data processing and application of InSAR crustal deformation observation. In this paper, the data processing and application of deep learning in InSAR crustal deformation observation are introduced, and its application prospect is prospected.

Key words: Deep learning, InSAR, Crustal deformation, Research Progress

中图分类号: