Features extraction for solar radio fine structures using fusion of wavelet and BEMD

In the traditional methods, the WT (Wavelet Transform) is used to extract fine structures of solar flares. However, the effect of the WT depends on the number of decomposition layers and the wavelet basis. The characteristics of different regions in the solar radio spectrogram are different. Correspondingly, the number of decomposition layers and the basis of the WT are also different. This results in the removal of few solar radio fine structures. To solve the above problems, considering the fact that the BEMD (Bidimensional Empirical Mode Decomposition) is based on original data, we combined the BEMD with the WT to extract the characteristics of the solar radio fine structures. Firstly, the WT and the BEMD are used to extract the high frequency components of the solar radio spectrogram as the solar radio fine structures exist in the high frequency components. And then, the high frequency reconstructed spectrograms are fused and cleaned to obtain a fine structure image with high quality. Finally, all fine structures in the whole burst area are determined by using the contiguous search algorithm. Simulation results show that compared with the WT, the method of extracting fine structures using fusion of wavelet and BEMD preserves the integrity of the solar radio fine structures better.