An Improved Wavelet Denoising Algorithm for Wideband Radar Targets Detection

A novel denoising technique based on wavelet transform modulus maxima (WTMM) is proposed for processing wideband radar spread targets detection signal in a clutter environment. Combined with the improved adaptive Bayes-Shrink threshold and Lipschitz exponents, we propose the path pruned approach at each scale terms as full-scale to split the signal. The estimation of WTMM over each scale has been optimized, thus, the signal and the noise can be split effectively. Additionally, to improve the computational efficiency, a fast method based on a piecewise polynomial interpolation algorithm is applied for the split signal reconstruction. Statistical results are quite promising and perform better than the conventional denoising algorithms: compared with the classical WTMM algorithm, the improved WTMM full-scale denoising algorithm not only increases the signal-to-noise (SNR) ratio by over 10 % but also reduces the processing time by 88 % and reduces the root-mean-square-error (RMSE) by over 35 %. More generally, the proposed algorithm has better performance than that of several typical algorithms in its denoising quality and singularity detection.