Estimation of Threshold Values at Various Scales

In partial discharge testing, acquired signals often contain white noise and periodic interference that need removal. This implementation utilizes the commonly used db6 wavelet from the Daubechies series to perform a 9-level multiresolution decomposition. Based on the energy characteristics of white noise, threshold values for each scale are estimated using hard thresholding processing, followed by signal reconstruction. The algorithm involves wavelet decomposition, noise variance estimation, and threshold application using MATLAB's wdenoise function or custom implementation with wthresh.