基于匹配追踪的谱分解方法及其应用

谱分解技术作为一项新颖的地震解释技术，在储层预测等方面得到越来越多的应用。准确刻画地震信号在时频空间分布特征是谱分解技术的关键。短时傅里叶变换的时频分辨率受时窗长度限制，不同的时窗参数对结果有较大的影响。小波变换以其多分辨率的特点能够得到更好的时频分布，但其在时间-尺度概念下的时频分解难以理解，而具有自适应特征的匹配追踪方法时频概念相对明确，模型和实际应用表明匹配追踪谱分解具有较好的时频分辨率，更适用于复杂的地震信号。

Spectral decomposition based on matching pursuit and its application

Spectral decomposition as a new technology of seismic interpretation has been adopted in more and more applications in reservoir prediction. The key to successful use of this technology is figure out a method which can accurately depict the time-frequency representation of seismic signal. The time-frequency resolution of short time Fourier transform is restricted by the width of the window. Window parameters have great impact on the results. Wavelet transform can give good time-frequency representation due ...