基于Gabor变换的地震资料高分辨率处理方法研究

High-resolution processing method of seismic data based on Gabor transform

 徐倩茹 孙成禹 乔志浩 唐杰

 （中国石油大学（华东）地球科学与技术学院，山东 青岛 266580）

 (School of Geosciences, China University of Petroleum, Qingdao 266580, China)

 日益复杂的勘探目标对地震资料的分辨率提出了更高的要求，常规的反褶积和反Q滤波等高分辨率处理方法往往需要精确的先验信息，同时忽略了低频成分的补偿，不能满足实际生产需要。文中采用基于时频变换的高分辨率处理方法，在Gabor时频变换的基础上，引入倒频谱。利用倒频谱的思想，在时频域对振幅谱进行对数改造，保持相位谱不变，同时采用平滑处理以补偿低频成分，联合实现高分辨率处理。将该算法应用到雷克子波、薄层、楔形体模型以及实际地震数据中，结果显示子波持续时间减小，低于调谐厚度的薄层得到识别，频带拓宽，分辨率得到了提高，验证了该方法的有效性。最后，探究了时窗宽度对处理结果的影响，结果表明，为了达到更好的高分辨率处理效果，Gabor变换的高斯窗宽度应不低于子波的持续时间。

 The increasingly complex exploration targets put forward higher requirement for the resolution of seismic data. Conventional high resolution processing methods, such as deconvolution and inverse Q filter, need accurate prior information, and these methods tend to ignore the low frequency component, which cannot meet the needs of the practical production. This paper adopts the high resolution processing method based on time-frequency transform and introduces the cepstrum on the basis of Gabor transform. In the time-frequency domain, we replace the amplitude spectrum with its logarithm, and remain the phase spectrum unchanged. At the same time, we use smoothing method to compensate for low frequency component to realize high resolution processing eventually. Applying this method to wavelet, thin layer, wedge model and real seismic data, the results indicate that this method can reduce the duration of wavelet, recognize the thin layer with thickness less than tuning thickness, broaden the bandwidth, and enhance the resolution. Finally this paper explores the impact of time window width on the processing result, which indicates that Gauss window width of Gabor transform should be not less than the duration of wavelet in order to realize high resolution processing.

10.6056/dkyqt201604012