数据与计算发展前沿 ›› 2025, Vol. 7 ›› Issue (4): 33-41.

CSTR: 32002.14.jfdc.CN10-1649/TP.2025.04.003

doi: 10.11871/jfdc.issn.2096-742X.2025.04.003

• 专刊：空间科学大数据智能算法模型与工具 • 上一篇    下一篇

电离层行进式扰动的自动识别与参数提取

赖昌1,2,*(),刘胜雨1   

  1. 1.重庆邮电大学， 电子科学与工程学院， 重庆 400065
    2.中国科学院国家空间科学中心， 北京 100190
  • 收稿日期:2025-05-07 出版日期:2025-08-20 发布日期:2025-08-21
  • 通讯作者: 赖昌
  • 作者简介:赖昌，博士，教授，重庆邮电大学，中国空间科学学会“空间科学大数据专业委员会”委员。主要研究方向为机器学习与中高层大气的交叉融合研究。
    本文中负责论文的架构、撰写与代码开发。
    LAI Chang, Ph.D., is a professor at Chongqing University of Posts and Telecommunications, CQUPT, and a committee member of the Big Space Science Data Professional Committee under the Chinese Society of Space Science. His main research interests include the intersection of machine learning and middle/upper atmospheric science.
    In this paper, he is responsible for the architectural design, manuscript writing, and code development.E-mail: laichang@cqupt.edu.cn
  • 基金资助:
    国家重点研发计划“空间科学大数据智能管理与分析挖掘关键技术及应用”(2022YFF0711400);中国科学院网信专项“空间天气典型事件知识挖掘与智能建模研究”项目(CAS-WX2022SF-0103)

Automatic Detection and Parameter Extraction of Medium-Scale Traveling Ionospheric Disturbances

LAI Chang1,2,*(),LIU Shengyu1   

  1. 1. School of Optoelectronic Engineering, Chongqing University of Posts and Telecommunications, Chongqing 400065, China
    2. National Space Science Center, Chinese Academy of Sciences, Beijing 100190, China
  • Received:2025-05-07 Online:2025-08-20 Published:2025-08-21
  • Contact: LAI Chang

RichHTML

0

PDF

7

摘要：

【目的】为解决中尺度行进式电离层扰动（MSTID）人工检测中存在的效率低、主观性强等问题，本研究提出了一种基于深度学习的三级处理架构，利用子午工程兴隆站点的氧原子气辉数据，实现MSTID的自动检测与参数提取。【方法】首先构建10层卷积神经网络模型对原始气辉图像进行环境分类，筛选出有效观测图像；其次，基于迁移学习策略和虚拟数据增强技术，训练快速区域卷积神经网络模型实现MSTID波面精准定位；最后通过边缘检测与线性拟合算法提取波动参数。创新性提出波面模拟函数与泊松-高斯混合噪声模型，生成虚拟训练数据以增强模型鲁棒性。【结果】分类模型在测试集上的准确率达到96.9%，检测模型的交并比普遍高于75%。本文开发的自动识别和参数提取系统显著提升了气辉数据处理自动化水平，为大规模电离层扰动统计研究提供了可靠的技术方案。

关键词: 中尺度行进式电离层扰动, 卷积神经网络, 快速区域卷积神经网络, 全天空气辉成像仪, 虚拟图像

Abstract:

[Objective] To address the inefficiency and subjectivity of traditional manual identification for medium-scale traveling ionospheric disturbances (MSTIDs), this study proposes an automatic deep learning-based framework for MSTID detection and parameter extraction from OI airglow images based on the data collected by the all-sky airglow imager at the Meridian Project’s Xinglong station. [Methods] The framework adopts a three-stage architecture: (1) A custom 10-layer convolutional neural network (CNN) is designed to classify raw airglow images, filtering out environmental interference (e.g., clouds, twilight overexposure) and retaining suitable observational data captured at starring night; (2) A Faster Region-based Convolutional Neural Network (Faster R-CNN) model, trained via transfer learning and virtual data augmentation, achieves precise localization of MSTID waveforms; (3) Edge detection and linear fitting algorithms are developed to extract propagation parameters such as direction, velocity, and wavelength. Innovatively, a hybrid Poisson-Gaussian noise model and a waveform simulation function are proposed to generate synthetic training data, enhancing model robustness against limited real-world samples. [Results] Evaluated on the test dataset, the framework demonstrates high performance: the CNN classifier attains 96.9% accuracy in identifying clear-sky conditions; the Faster R-CNN detector achieves an average Intersection-over-Union (IoU) of >75% for wavefront localization. The proposed system significantly improves the automation and objectivity of MSTID analysis, enabling efficient large-scale statistical studies of ionospheric disturbances.

Key words: medium-scale traveling ionospheric disturbances, CNN, Faster R-CNN, all-sky airglow imager, synthetic image