Two Dimensional Empirical Mode Decomposition and Hilbert Spectral Analysis for Image (GSC-13909)

Abstract:
Traditional energy-frequency analyses are based on Fourier transformation. These methods produce consistently accurate results only for linear and stationary signals. However, few energy-frequency data sets are truly linear and stationary. Although fast Fourier transformation (FFT) is routinely used for these nonlinear, nonstationary signals, results are often inaccurate. NASA Goddard Space Flight Center has developed a new tool specifically designed for analysis of nonlinear, nonstationary data. The Hilbert-Huang transformation algorithms accurately analyze energy frequencies via a three-step process: * Sifting: Based on the empirical mode decomposition method, HHT¿s sifting process generates a collection of simple mode functions (SMFs) for the complicated data, allowing instantaneous frequencies to be defined. * Hilbert transformation: Performing a Hilbert spectral analysis on the SMFs converts the energy-based signals to frequency- and time-based signals. * Spectral presentation: The results are presented within the context of the Hilbert spectra. Although designed to analyze nonlinear, nonstationary data, the innovative HHT method also can be used for data that are linear and stationary, delivering results that are identical to FFT analysis HHT algorithms are coded in C++, and when compiled they can be executed on any PC or workstation. The operation does require the use of 32-byte addresses, and therefore a Windows 95 or Windows NT operating environment is required.

Type of Offer: Licensing



Next Patent »
« More Physics Patents

Share on      


CrowdSell Your Patent