Precision Time-frequency Analysis And Application

Time-Frequency (TF) analysis is very important in signal processing. The main mission of TF analysis is to describe the variety of the frequency followed time, and gain one kind of time-frequency energy distribution which express the signal both time domain and frequency domain. In time-frequency domain, that distribution can show clearly the characteristic of a signal which can't be captured in time domain. Using time-frequency energy distribution we can abstract information in given time or frequency. So TF analysis had been one main tool in signal processing. Nevertheless past or present TF analysis methods had various deficiencies which could not be conquered by them. These shortcomings will cause it's difficult to interpret the result or need abundant mathematic illations.The Hilbert-Huang Transform (HHT) is a signal analysis theory, which developed through integrate novel EMD and classical Hilbert transform. It can express the intrinsic essence using simple and understandable algorithm. On other hand, the analysis result achieved excellence precision. Hilbert spectrum is direct time-frequency energy distribution. The points on time-frequency grids described meaningful and actual distribution. In HHT theory, Innovations were brought up in signal decomposition, instantaneous parameter computation and time-frequency analyses. In this paper, we will review the most important methods in signal processing, and introduce the algorithm of EMD and Hilbert transform, then indicate principle of Hilbert-Huang transform and its peculiarity.Additionally, in this paper, some improvements will be provided in interpolation methods, dealing boundaries and terminate condition. Compare the effect of HHT and others means in speech signal and some model data application. And apply the HHT method to seismic signal processing because of its advantage.The research in this paper made some efforts in supplement HHT theory and application effect of EMD, and provided new methods for seismic data process, speech analysis and edge detection. At the same time, this paper also provided abundant experience of research on HHT application.