Automated Detection and Statistical Study of Solar Radio Spikes by P.R. Lv et al

Solar radio spikes’ most typical observational features are their short duration and narrow bandwidth. They appear on the solar radio dynamic spectrogram as a large number of narrow-band type-III bursts, spikes, dots, sub-second clumps, groups, chains, and other narrow-band structures from decimeter to decameter wavelengths (Feng et al.(2018), Tan et al.(2019)). We have improved the YOLOv5s network model for these characteristics by adding inclined bounding frames and attention and feature fusion mechanism modules. Recently, Hou et al. (2020) identified and extracted the solar radio spikes using the Faster Region-based Convolutional Neutral Network (Faster R-CNN), giving the AP value 91%.

The decimeter- and meter-wavelength spikes observed by the Solar Broad-band Radio Spectrometer in Huairou and the Chashan Solar Radio Observatory spectrograph are used to conduct experiments, respectively. The spikes at decimeter and meter wavelengths are again categorized based on their frequency-drift rates, such as positive, negative, and no measurable frequency-drift rates. We have carried out a statistical study on these categorized spikes. These statistical results and findings constrain solar radio spikes’ formation.

Figure 1. (a) Precision-Recall relationship. (b) In comparing labeling and detecting, the left column is two images with labeled spikes, and the right is images with detected points from the improved network.

The relationship between Precision and Recall is shown in Figure 1(a). Precision is TP/(TP+FP), representing what proportion of the target detected by the model is real, that is, the real spikes. Among them, the Recall is equal to TP/(TP+FN), which means what proportion of the real target can be detected by the model. In the formula, TP, FP, and FN are the spikes numbers with positive predictions and labels, positive predictions but negative labels, and negative predictions but positive labels, respectively. In Figure 1(a), when Recall is 0.64, Precision equals 0.85. Once detect the inclined bounding frames of spikes are present, it is necessary to evaluate and analyze the effectiveness of the detection. AP value is used as an evaluation indicator. It’s determined by the Precision-Recall curve and is equal to the area enclosed by the curve. The datasets are input into the original and improved YOLOv5s network2. The models are trained according to the official pre-training weights file of YOLOv5s.pt3 with a training time of 1200. The detection results from these two models are compared.

Figure 1(b) shows a comparison of the detection results. The figure contains four sub-images. The two left columns show the labeled spikes in datasets, and the right two columns are detected spikes by the improved network. The improved network performed well regarding small targets detection accuracy and inclined bounding frames regression precision.

The YOLOv5s network is improved concerning the typical characteristics of spikes. In the improved network, inclined bounding frames are added to detect their frequencies drifting with time. Given various morphologies, short durations, and narrow bandwidths, attention and feature fusion mechanism modules are added. The decimeter-wavelength and meter-wavelength spikes observed by the SBRS/Huairou and the spectrograph in the CSO are used to conduct experiments, respectively. The results demonstrate that the AP value obtained by the improved network is 74%. The value is almost 14% higher than that of the original network.

The improved network detects 9709 (1379) decimeter- (meter-) wavelength spikes in two events that occurred on 2005 January 20 (2016 July 18) with durations, bandwidths, relative bandwidths, and frequency-drift rates. According to the frequency ranges and frequency-drift directions, the spikes are classified.

Figure 2. Histograms of spikes at decimeter (the left column) and meter (the right column) wavelengths. The first, second, and third rows are the duration, the bandwidth, and the relative bandwidth distributions, respectively. The curves are from the Gaussian fitting.

In order to analyze the distribution of spikes, we plot histograms of their durations, bandwidths, and relative bandwidths, as shown in Figure 2. The figure shows that at the decimeter wavelengths, the durations with a good symmetrical distribution mainly lie in 11-12 ms, and the numbers larger and smaller than this period are similar. The bandwidths and relative bandwidths show apparent asymmetry, primarily at 20 MHz and 2%, respectively. Their numbers are smaller than the two values, and those more than them are larger. At the metric wavelengths, the durations also show an excellent symmetrical distribution, and most of the spikes are within 40-50 ms. The values are not equally distributed for the bandwidths and the relative bandwidths. We carry out a statistical study on the categorized spikes and find the following main results:

(1) The duration declines with the increase of frequency, the bandwidth increases with frequency and declines with duration, and the relative bandwidth remains unchanged with frequency and duration. The duration (bandwidth) at the decimeter wavelengths is about one-quarter (4-5 times) of that at the meter wavelengths. The decimeter (meter) wavelength duration has an excellent symmetrical distribution and mainly lies in 11-12(40-50) ms. The bandwidth and relative bandwidth at the two wavelengths show asymmetrical distributions.

(2) At the decimeter wavelengths, spikes with no measurable frequency-drift rates have the most significant number, the negative is the second, and the positive is the last. The mean duration (11.2ms) and bandwidth (21.8MHz) of no measurable frequency-drift spikes are slightly smaller than the means of positive (12.1ms and 25.1MHz) and negative (12.2ms and 23.7MHz) frequency-drift spikes. At the meter wavelengths, spikes with positive and negative frequency-drift rates are almost similar, and no measurable frequency-drift rates are the least.

Based on the recent paper by Lv, P.R., Hou, Y.C., Feng, S.W. et al. Automated detection and statistical study of solar radio spikes, Astrophys Space Sci 368, 14 (2023). DOI: 10.1007/s10509-023-04172-8


Feng, S.W., Chen, Y., Li, C.Y. et al. Harmonics of Solar Radio Spikes at Metric Wavelengths. Sol Phys 293, 39 (2018).

Hou, Y.C., Zhang, Q.M., Feng, S.W. et al. Identification and Extraction of Solar Radio Spikes Based on Deep Learning. Sol Phys 295, 146 (2020).

Tan B, Chen N, Yang Y H, et al. Solar Fast-drifting Radio Bursts in an X1.3 Flare on 2014 April 25[J]. The Astrophysical Journal, 2019, 885(1):90.

*Full list of authors: P.R. Lv, Y.C. Hou, S.W. Feng, Q.F. Du and C.M. Tan