On the local GST power spectrum distribution of color noise
Jinghuai Gao, Weishi Man, Wen-chao Chen, Shumin Chen · 2004
The S transform (ST) and generalized S transform (GST) are new tools for signal processing. In this work, the mean ST and GST power spectrum of white noise are derived, respectively. The results show that both are linear with frequency. And then local GST power spectrum of color noise is investigated by qualitative analysis, it is concluded that the local GST power spectrum follows chi-square distribution with two degree of freedoms. These conclusions are further certified by Monte Carlo method. Based on the above results, a method distinguishing useful signals from noise is proposed in GST domain. The effectiveness of this method is proved by synthetic data