It is important to improve seismic resolution for thin reservoir characterization.Conventional methods,e.g.deconvolution and inverse Q filtering,are not effective enough owing to the problems of rigorous assumptions,low signal-to-noise ratio,and low robustness.A new method based on matching pursuit algorithm was proposed,which can improve the seismic resolution while maintaining high signal-to-noise ratio.Firstly,according to the characteristics of seismic signals,a wavelet dictionary was construc-ted using a frequency-weighted-exponential(FWE)function,which performed well in fitting spectra of different shapes.Then,seis-mic data were decomposed as the linear superposition of the wavelets in the wavelet dictionary using the matching pursuit algo-rithm,followed by amplitude enhancement for those matched wavelets with high frequencies and high signal-to-noise ratios.Final-ly,the matched wavelets were summed up to obtain high-resolution seismic data.The application test showed that processed data with high resolution and high signal-to-noise ratio agreed with synthetic seismogram calculated using log data.Reliable results will facilitate subsequent thin reservoir characterization.
matching pursuithigh resolutionseismic data decompositionwavelet dictionarythin bed
维普
万方数据
