论文部分内容阅读
分析了周期混叠的来源及其对频域正演波形记录精度的影响.在采样间隔固定时,快速傅里叶变换(FFT)变换不同的抽样点数N对正演波形记录精度有很大影响,提出了一种改善周期混叠的方法,通过合理地增大FFT变换的抽样点数使得单位冲激响应在N点处幅值更接近零,进而降低各个周期的混叠程度以提高频域波形正演和反演的精度.基于复杂模型Marmousi模型的正演和反演实验结果表明,该方法能较大幅度降低周期混叠程度,有效地提升正演和反演结果精度.在理论数据长度固定时,抽样点数过大反而会降低反演精度.本文以Marmousi模型原始观测系统以及数据长度为例,说明可以选取最佳的抽样点数来降低周期混叠对频域波形正演和反演的影响.
The source of periodic aliasing and its effect on the recording accuracy of the forward waveform in the frequency domain are analyzed.When the sampling interval is fixed, the number of sampling points N in the FFT transform has a great influence on the recording accuracy of the forward waveform , A method of improving the period aliasing is proposed. By reasonably increasing the sampling points of the FFT transform, the amplitude of the unit impulse response is closer to zero at the N point, thereby reducing the aliasing degree of each period to improve the frequency domain waveform Forward modeling and inversion accuracy.The results of forward modeling and inversion based on the complex model Marmousi model show that this method can greatly reduce the degree of periodic aliasing and effectively improve the accuracy of forward modeling and inversion results.With the theoretical data length Fixed sampling point is too large but will reduce the accuracy of inversion.In this paper, the original observation system Marmousi model and data length as an example, we can select the best sampling points to reduce the periodic aliasing on the frequency domain waveform inversion and forward influences.