Use of Threshold and No-Threshold Methods of Discrete Wavelet Filtering of Radar Signals

The efficiency of threshold and no-threshold wavelet noise filtering methods in processing radar signals is analyzed. Three methods of wavelet analysis are considered: a method with a general limit threshold for wavelet coefficients of detail; a method with a universal threshold for all wavelet decompositions; and a method without a threshold, based on zeroing the coefficients of detail at certain levels of the wavelet decomposition of the signal. The efficiency of wavelet filtering by the signal-to-noise ratio before and after filtering, signal entropy, and the model’s mean squared error (MSE) are evaluated. It is found that over a wide range of high noise from –12 dB to –7.5 dB, the common-threshold method provides more efficient noise filtering than the other methods.