Application of Deep Learning to Recognize Ionograms

Mochalov, Vladimir; Mochalova, Anastasia

doi:10.1109/rwp.2019.8810326

Cited by 8 publications

(6 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As a comparison, the Dice-Coefficient Loss (DCL) values achieved by Mochalov and Mochalova (2019) [ 7 ] are 0.18859, 0.22881, and 0.22209 for F2, F1, and E layers, respectively, for ionograms compressed to a size of 192 × 144 pixels. These values can be converted to IoU by Equations (5) and (6):

where X is ground truth and Y is the prediction by models.…”

Section: Resultsmentioning

confidence: 99%

“…Based on the same calculation, the recall rate and F-score achieved by our SA-UNet model are 0.982 and 0.9339. The best predictions of F2 from Mochalov and Mochalova [ 7 ] and Xiao et al, [ 8 ] are compared with our best prediction in Table 6 .…”

Section: Resultsmentioning

confidence: 99%

“…They used intersection over union (IoU) to evaluate the performance of their model and reported an average IoU of 0.569. Mochalov et al, 2019 [ 7 ] applied a deep neural network (DNN) U-Net to identify E, F1, and F2 layers of mid-latitude ionosphere observed by the ionosonde “Parus-A” at Kamchatka. The three layers selected for identification in this work are strong and do not overlap with each other in their ionograms.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

State-of-the-Art Capability of Convolutional Neural Networks to Distinguish the Signal in the Ionosphere

Chang

Lin

Дмитриев

et al. 2022

Sensors

View full text Add to dashboard Cite

Recovering and distinguishing different ionospheric layers and signals usually requires slow and complicated procedures. In this work, we construct and train five convolutional neural network (CNN) models: DeepLab, fully convolutional DenseNet24 (FC-DenseNet24), deep watershed transform (DWT), Mask R-CNN, and spatial attention-UNet (SA-UNet) for the recovery of ionograms. The performance of the models is evaluated by intersection over union (IoU). We collect and manually label 6131 ionograms, which are acquired from a low-latitude ionosonde in Taiwan. These ionograms are contaminated by strong quasi-static noise, with an average signal-to-noise ratio (SNR) equal to 1.4. Applying the five models to these noisy ionograms, we show that the models can recover useful signals with IoU > 0.6. The highest accuracy is achieved by SA-UNet. For signals with less than 15% of samples in the data set, they can be recovered by Mask R-CNN to some degree (IoU > 0.2). In addition to the number of samples, we identify and examine the effects of three factors: (1) SNR, (2) shape of signal, (3) overlapping of signals on the recovery accuracy of different models. Our results indicate that FC-DenseNet24, DWT, Mask R-CNN and SA-UNet are capable of identifying signals from very noisy ionograms (SNR < 1.4), overlapping signals can be well identified by DWT, Mask R-CNN and SA-UNet, and that more elongated signals are better identified by all models.

show abstract

where X is ground truth and Y is the prediction by models.…”

Section: Resultsmentioning

confidence: 99%

Section: Resultsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

State-of-the-Art Capability of Convolutional Neural Networks to Distinguish the Signal in the Ionosphere

Chang

Lin

Дмитриев

et al. 2022

Sensors

View full text Add to dashboard Cite

show abstract

“…Mochalov & Mochalova, 2019 detect the E, F1 and F2 layers of the ionosphere respectively by U‐Net (Ronneberger et al., 2015). Jara & Olivares, 2021 propose a semi‐supervised learning neural networks model based on the multi‐convolution decoder which is superior to the contour detection method based on image processing, threshold method and unsupervised machine learning.…”

Section: Related Workmentioning

confidence: 99%

“…However, there are few types of research on ionospheric echo extraction based on deep learning technology. In terms of processing the vertical ionograms, deep learning (Jara & Olivares, 2021; Mochalov & Mochalova, 2019; Xiao et al., 2020) is used to scale ionograms and detect echoes, and the performance of these methods is better than traditional methods. However, there are few reports on oblique and backscatter ionospheric echo extraction.…”

Section: Introductionmentioning

confidence: 99%

Ionogram Echo Extraction Based on the Convolutional Neural Networks

et al. 2022

View full text Add to dashboard Cite

Free electrons and ions in the ionosphere can change the propagation speed and direction of radio waves, which greatly affects the performance of wireless communication, navigation, radar positioning, and so on (Ravan et al., 2012;Zhou et al., 2011). The effective extraction of ionospheric echoes is beneficial to the research of ionospheric characteristics, for example, the physical process, the characteristics of the ionospheric state, and the law of ionospheric radio wave propagation. In addition, it supports the study of radio transmission theory and provides a theoretical basis for the design, use and detection of the radio engineering system.Recently, some studies on the extraction of ionospheric echoes mainly focus on three types of ionograms, which are obtained by vertical, oblique and backscatter sounding (

show abstract

Deep-learning for ionogram automatic scaling

Xiao

Wang

et al. 2020

Advances in Space Research

View full text Add to dashboard Cite

Application of Deep Learning to Recognize Ionograms

Cited by 8 publications

References 5 publications

State-of-the-Art Capability of Convolutional Neural Networks to Distinguish the Signal in the Ionosphere

State-of-the-Art Capability of Convolutional Neural Networks to Distinguish the Signal in the Ionosphere

Ionogram Echo Extraction Based on the Convolutional Neural Networks

Deep-learning for ionogram automatic scaling

Contact Info

Product

Resources

About