习宇

摘      要：采样率是地震数据处理中的一个重要参数。在反褶积时采样率依赖原始地震数据的采样率，采样率越小，反褶积效果的精度就越高，但野外采集花费的成本和室内数据处理所用的时间也越多，而采样率越大，虽能节约成本和提高效率，但反褶积效果的精度会变低。此外，信号的频谱在理论上是无限宽的，采样将丢失一部分信息，这就可能会对反褶积效果造成影响。针对以上问题，本文基于采样定理研究采样率对反褶积效果的影响，通过对地震数据重采样，对重采样后的地震数据进行Gabor反褶积处理。数值模拟和实际资料结果表明：在遵循采样定理的情况下，大采样率的反褶积效果无法识别薄层;从反褶积效果的精度和地震资料处理效率来看，存在一个最优采样率，既能保证反褶积效果的精度，又能提高工作效率。

关  键  词：采样率;采样定理; 频谱;Gabor反褶积

中图分类号：TE19       文献标识码： A       文章编号： 1671-0460（2020）04-0724-04

Abstract：  Sampling rate is an important parameter in seismic data processing. The sampling rate at deconvolution depends on the sampling rate of the original seismic data.The smaller the sampling rate， the higher the accuracy of the deconvolution effect. However， the cost of field collection and the time spent on indoor data processing are also increased. The large sampling rate saves the cost and increases the efficiency， but the accuracy of the deconvolution effect becomes lower. In addition， the spectrum of the signal is theoretically infinitely wide， and the sampling will lose part of the information， which may affect the deconvolution effect. In view of the above problems， the influence of sampling rate on the deconvolution effect was studied based on the sampling theorem. By resampling the seismic data， the resampled seismic data were processed by Gabor deconvolution. The numerical simulation and real seismic data showed that the deconvolution effect of large sampling rate could not identify the thin layer under the condition of sampling theorem; from the accuracy of deconvolution effect and seismic data processing efficiency， there was an optimal sampling rate， which could not only ensure the accuracy of the deconvolution effect， but also enhance work efficiency.

Key words： sampling rate; sampling theorem; spectrum; Gabor deconvolution

反褶積的主要目的是提高分辨率，菲涅尔带对于研究横向分辨率具有重要意义[1]，但分辨率与信噪比在某种程度上是相互制约的，反褶积后信噪比的降低取决于反子波能量的大小[2]。常规反褶积方法基于反射系数序列为白噪，而实际反射系数序列呈现蓝谱特征[3， 4]，谱模拟反褶积[5-9]能够对非白噪反射系数序列有更好的适应性，该方法是在假设子波振幅谱光滑的情况下，利用数学方法将子波振幅谱提取出来。谱模拟反褶积方法能够很好地适应于子波时不变的情况，而地震子波在地下传播过程中，能量衰减，表现出时变特征，造成地层深部信息丢失。……