【摘要】 小波多尺度分析可以有效处理一维信号的点奇异特征,但对于二维信号的线奇异特征,小波变换显得无能为力。Curvelet多尺度变换可以对时空信号进行最稀疏表达,能够获得最优的非线性逼近。通过分析地震信号在Curvelet域三维空间的特征,认为时空信号的不同波组成分在Curvelet域存在明显的差异,可以从频率、角度和空间位置实现有效反射波和干扰波的分离。理论模型与实际单炮记录处理结果表明,Curvelet域方法在分离干扰波、突出反射波的同时,可以较好地保持有效波信息,保真度好。更多还原

【Abstract】 The multi-scale analysis can effectively describe the signal characteristics in different scales efficiently and is widely used in signal processing.Curvelet transform is characterized by optimum sparseness constraint condition that can deal with line-like phenomena in high dimension.Through analyzing the signal characteristic in Curvelet domain among different waves,we think they have significant difference in frequency,angle and position.So we can employ Curvelet transform to separate interference wave from effective wave.Examples of processing the model data and field data illustrate its feasibility and effectiveness in wave separation,it can provide superior interference wave attenuation with minimal impact on the desirable signal components,and its result has good fidelity.更多还原