Research on Target Micromotion Features Extraction Based on Wavelet Transform and Hilbert Transform

Identification of low and slow small targets in multi-band radar requires high-precision micromotion features, while traditional data analysis provides methods with features that are not obvious and difficult to quantify. Aiming at this problem, two methods based on distance dimension transform are proposed to extract features from radar echo data, and the extracted parameters are compared and analyzed. The results show that the feature extraction accuracy of the Hilbert algorithm is better than that of the wavelet transform, and its information entropy is reduced by 6.31 bits, reduced by 91.31% compared to the wavelet transform; the signal-to-noise ratio is improved by 42.55 dB, which is about 10 times higher than that of the wavelet transform. In addition, these two feature extraction methods have the highest feature extraction accuracy for DJI Inspire 2 DJI Matrice 600 UAVs, which have been implemented in recent years, and their signal-to-noise ratios can reach up to 50.39dB.