Due to the limitation of low-resolution radar system and the influence of background clutter in the detection process, it is hard for low-resolution radars to classify and identify aircraft targets. To solve the above problems, a classification method for aircraft based on Ensemble Empirical Mode Decomposition (EEMD) and multifractal is proposed, in which the intrinsic modes are obtained by EEMD, and the waveform entropy in the Doppler domain is used to screen and reconstruct the intrinsic modes. The multifractal feature of the target echo data is extracted from the reconstructed signal, and then the aircraft target classification and recognition experiment is carried out with support vector machine. The experimental results show that the feature data extracted by ensemble empirical mode decomposition and multifractal analysis can be used for the classification and identification of civil aircraft and fighter aircraft, and the accuracy rate is about 98.5%, which is higher than that of timedomain multifractal method.

No abstract

The research goal of low-resolution radar aircraft target classification is to analyze the category of the given low-resolution radar aircraft target echo. In existing solutions, the feature extraction methods based on rotating modulation spectrum have good performance, such as the complex cepstrum method, autocorrelation method, cycle diagram method, autoregressive model power spectrum method, and singular value decomposition method. Most of these methods are more complicated in calculations, and practical applications often require higher pulse frequencies and longer observation times, which are greatly restricted. In this paper, a classification method based on ensemble empirical mode decomposition and multifractal correlation (CMEEMDMFC) is proposed. The basic design idea is to obtain the intrinsic mode functions (IMFs) by using the signal decomposition ability of ensemble empirical mode decomposition (EEMD) and select some components which are beneficial for improving the signal-to-noise ratio (SNR) for recombination. Then extract the corresponding multifractal correlation (MFC) features from the new signals for recognition. For verifying the validity of the model, a comparison model was selected to test on the same data set. Experimental results show that the proposed model performs well in classification accuracy.

No abstract

No abstract