Abstract:

To improve the de-noising performance of filters, we present a new method to optimize the two thresholds based on particle swarm optimization (PSO) in the wavelet domain. Having analyzed the widely used soft threshold and hard threshold de-noising methods, we select the Donoho threshold as the upper threshold, and minimax threshold as the lower threshold. The PSO algorithm is used to optimize the double-threshold for the best threshold for de-noising. Simulation results show that the proposed method can overcome the psuedo-Gibbs phenomenon of the hard-thresholding method and excessive smoothness of the signal caused by soft-thresholding. It enhances SNR and reduces RMSE, providing better performance than the traditional methods.

Key words: wavelet de-noising, double-threshold, soft-threshold, hard-threshold, PSO