Research On Empirical Mode Decomposition Algorithm And Its Application In Electromagnetic Imaging

As an important application of electromagnetic imaging,radar imaging is playing an important role in military and civilian fields.As an important kind of radar imaging method,Inverse synthetic aperture radar(ISAR)has the ability of observing and locating the targets in all-weather and long distance.In classic ISAR imaging methods,imaging data is assumed stationary so that ISAR image could be obtained by Fourier Transform(FT).However,in practice the imaging target is often in motion,hence the Doppler frequency of moving target is usually time varying.In this case,the received signal is non-stationary and the resulting ISAR image by FT is blurred.One effective approach for this problem is to use time-frequency transform(TFT)to replace FT in the process.Wigner-Ville Distribution(WVD)is an efficient time-frequency transform with good time-frequency resolution for ISAR imaging.But for multi-component signals the cross-term interference of WVD may be serious.And it causes that the image is also not clear enough.As a powerful tool for the analysis of non-stationary signals,Empirical Mode Decomposition(EMD)can decompose multi-component signals into several Intrinsic Mode Functions(IMF)from high frequency to low frequency.And it is effective to reduce the cross-term interference of WVD of multi-component by using EMD.According to the characteristics of ISAR imaging data,the ISAR imaging methods based on CEMD-WVD are introduced.Then the ISAR imaging methods base on CEMD and instantaneous frequency are proposed.The two kinds of complex-valued Empirical Mode Decomposition method are introduced in this paper.They are Complex Empirical Mode Decomposition(CEMD)and Bivariate Empirical Mode Decomposition(BEMD).And the ability for decomposed multi-component of CEMD and BEMD is contrasted by the numerical simulation.Finally,the ISAR imaging methods based on BEMD and instantaneous frequency and the ISAR imaging methods based on BEMD-WVD are proposed.