Font Size: a A A

Higher Order Statistics Methods Of Digital Processing Of Seismic Data

Posted on:2006-02-25Degree:MasterType:Thesis
Country:ChinaCandidate:D W LiFull Text:PDF
GTID:2190360182456000Subject:Earth exploration and information technology
Abstract/Summary:PDF Full Text Request
High-order statistics analysis is in front of domestic and international signal processing realm nearly twenty years. High-order statistics applies to all problem that need consider non-gauss, non—minphase, non-linear and circular calm. In the seismic signal process realm, high-order statistics can be used to eliminate color noise, at the same time, it is a powerful tool for identification and estimation non-minphase wavelet. Beside, Independent component analysis is a new technique which base on the character of high-order statistics and developped in recent years. It has become the hotspot of signal process realm through more ten years continuously perfect.The paper search more on the foundation of the result research fore people on the side of seismic wavelet estimation and processing. At the same time, applying independent component analysis to eliminate noise of seismic data in a creative way. Completed primarily below work:1. Combining the distribute character of reflection coefficient to research trait of high-order statistics that every kind of distribute serial; The contrast analyzed their ability of resist noise towards the different orders through estimation high-order statistics of different type of wavelet; Analyzing character of high-order statistics that seismic trace, and bring forward and providing the relation between trace and its four-order statistics.2. Import magnitude reconstruct method to estimation wavelet base on high-order statistics after phase estimation theory, and perfecting the method of estimation wavelet though high-order statistics. Next, according reconstructed wavelet to deconvolution seismic data, this increase resolving power of seismic data.3. Applying independent component analysis to eliminate noise of seismic data according different distribution of valid signal and noise in seismic data. The feasibility ofthe method proved through simulation test and real data processing.
Keywords/Search Tags:High-order statistics, Seismic wavelet, Deconvolution, Independent component analysis, Kurtosis, Negentropy, Mutual information
PDF Full Text Request
Related items