Using wavelet-based indices for detecting abrupt changes in signals
Mohamed Allali, Victor DeBrunner · 2001
A new method based on the wavelet transform (WT) to detect abrupt changes in nonstationary noisy signals is presented. We show that the wavelet transform combined with the short time Fourier transform (STFT) provides a more accurate index of stationarity than the STFT alone, especially for very noisy signals.