【摘要】 时频分析技术,是研究非平稳信号的关键技术,提取与含油气相关的局部时频属性,为进一步寻找复杂油气藏提供重要依据.其中,匹配追踪分解是具有高时频聚焦性的时频分析技术,不受不确定性原理的约束.原子库与待分析信号结构的相关性是影响匹配追踪分解效果的关键因素.根据地震信号的特点,探讨了基于宽带Marr子波簇原子库的匹配追踪算法,宽带Marr子波与常规的Ricker子波相比,具有更丰富的波形特征,子波旁瓣大小具有可调控性,更能适应地震子波的结构特征,具有较强的自适应性.模型和实际应用都取得了较好的效果.更多还原

【Abstract】 Time-frequency analysis technology is a key method to study non-stationary signal.The technology helps to extract seismic attributes related to oil and gas,so as to provide important basis for detecting complex reservoirs. Matching tracking decomposition is a kind of time-frequency analysis technology with high time-frequency focusing property and free of uncertainty principle.The composition correlation between the dictionary of atoms and the signal which needs analyzing is the significant factor for the matching tracking decomposition.Based on the features of seismic signal,the matching tracking algorithm of wide band Marr wavelet cluster atom dictionary is discussed in this paper.Compared with Ricker wavelet,the Marr wavelet has more detailed and abundant waveform information,easily controlled wavelet side lobes,thus it has better self-adaption ability and better adopt to the real seismic wavelet.The synthetic and real data examples illustrate the effect of the technology.更多还原