A Deep Learning Approach for Automatic Ionogram Parameters Recognition With Convolutional Neural Networks
Typical ionosondes operate with >5 min time intervals, which is enough to obtain regular parameters of the ionosphere, but insufficient to observe short‐term processes in the Earth's ionosphere. The key point for this study is to increase the ionosondes data time resolution by automatization...
Gespeichert in:
Veröffentlicht in: | Earth and Space Science 2024-10, Vol.11 (10), p.n/a |
---|---|
Hauptverfasser: | , , , |
Format: | Artikel |
Sprache: | eng |
Schlagworte: | |
Online-Zugang: | Volltext |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
Zusammenfassung: | Typical ionosondes operate with >5 min time intervals, which is enough to obtain regular parameters of the ionosphere, but insufficient to observe short‐term processes in the Earth's ionosphere. The key point for this study is to increase the ionosondes data time resolution by automatization of ionogram scaling routine. In this study we show the results of implementation of deep learning approach for ionogram parameters scaling. We trained the model on 13 years ionogram data set of Sodankyla ionosonde at high latitude region (67°N). We tested our autoscaling program tool on 2021 years data set and evaluate errors between operator and automatic parameters scaling. The root mean square errors for critical frequencies foF2, foF1, foE, foEs, fmin, fbEs and virtual heights h′F, h′E, h′Es are estimated as 0.12 MHz (2 pixels), 0.07 MHz (1.16 pixels), 0.15 MHz (2.5 pixels), 0.33 MHz (5.5 pixels), 0.15 MHz (2.5 pixels), 0.17 MHz (2.83 pixels), 7.7 km (1.34 pixels), 7.0 km (1.22 pixels), 7.1 km (1.24 pixels), respectively.
Key Points
The deep learning approach was effectively deployed to scale ionograms to obtain ionospheric parameters in the polar region ionosphere
The performance of common convolutional neural network models was enhanced by implementing a noise‐to‐noise approach for ionogram pre‐filtration, incorporating data augmentation techniques, and diversifying the data set
The proposed deep learning general architecture demonstrates consistent performance on independent test data set |
---|---|
ISSN: | 2333-5084 2333-5084 |
DOI: | 10.1029/2023EA003446 |