地震信号建模与应用

勘探地震学中精确的反演成像是核心问题。地震数据的预处理对反演成像结果起着至关重要的作用。在现代信号分析理论下对地震信号进行分析和处理是必然的发展方向，其中信号模型的建立和Bayes框架下的稀疏和冗余的信号表达是关键。地震信号是特定震源激发在地层介质中传播并观测到的反射/散射波场，线性和非线性时间相位规定的地震反射子波加上随机噪音是地震信号的基本描述。基于上述总的判断，本项研究试图从总结实际地震数据中地震信号和噪音特点出发，根据地震波散射理论分析带限信号子波的波形特征，建立描述子波的参数模型；借助统计方法建立统一的包含线性和非线性时间相位变化特点地震信号模型，发展实用有效的信号稀疏和冗余表达的计算方法。所发展的模型示范性地应用于高维地震数据规则化和噪音压制中。

Accurate inversion imaging is the core of exploration seismology. Seismic data preprocesing is crucial for final imgaging quality. It is consequent to analysis and process seismic signal under the theory of modern signal analysis,among this the key is seismic model building and sparse and redundant representating under Bayes estimation. Seismic signal is a reflected and scattered wavefield which is generated by a certain manual source and propagates in the geological meidum.The wavelets restricted by linear and/or nonlinear time-phase and the random noise is the basic description of the seismic data.Based on this judgement, this research proposal will follow the route. Firstly, the characteristics of seismic data is summarised, and then the parameter model of the wavelet is bulit through the diffraction wave theory. Secondly, linear and nonlinear signal model is bulit with the help of the statistical methods and the numerical methods about the sparse and redundant representation is developed.Finally,the developed signal models are exemplarily used in the high-dimensional seismic data regularization and noise attenuation.