P1.25
Denoising atmospheric turbulence signal with wavelet-packet transform
Hu Fei, Chinese Academy of Sciences, Beijing, China; and Y. Jing
In this paper, the idea of the wavelet packet was introduced, the state-of-the-art tool in analyzing various signals, especially about the original wavelet packet. And then the author de-noised the turbulence data with the wavelet packet by the global threshold proposed by Donoho and Johnstone. Which can de-noise the data with great accuracy, sometimes even when you do not know the type of the noise, and use this method, the spectrum characteristic become quite well. It can be concluded that wavelet packet transform made possible a new method for extricating signals from white noise (“all-color”, or all frequency noise), which has great potential in many fields including experimental atmospheric turbulence data analysis.
Poster Session 1, Posters for the 26th Conference on Agricultural and Forest Meteorology
Wednesday, 25 August 2004, 5:30 PM-8:30 PM
Previous paper Next paper