基于PSO的地震盲反褶积方法  

A blind seismic deconvolution method based on Particle Swarm Optimization

作　　者：蔡连芳 田学民[1] 

机构地区：[1]中国石油大学(华东)信息与控制工程学院,青岛266580

出　　处：《地球物理学进展》2012年第3期1116-1122,共7页Progress in Geophysics

基　　金：国家863计划项目(编号:2007AA04Z193);山东省自然科学基金资助项目(编号:Y2007G49)

摘　　要：本文基于地层反射系数非高斯的统计特性,在反褶积输出单位方差约束下,将反褶积输出的负熵表示为非多项式函数,作为盲反褶积的目标函数,然后采用粒子群算法优化目标函数寻找最佳反褶积算子,实现地震信号的盲反褶积.数值模拟和实际资料处理结果表明,与传统反褶积方法相比,本文方法同时适应于最小相位子波及混合相位子波的反褶积,能够更好地从地震数据中估计反射系数,有效拓宽地震资料的频谱,得到高分辨率的地震资料.The deconvolution of seismic data is to estimate reflectivity so that the resolution can be improved. In most cases, reflectivity exhibits the non-Gaussian statistical character. The information theory criterion--negentropy is a good measure tool of non-Gaussian character. When the negentropy of deconvolution output attains maximum, deconvolution output can be thought as the optimal estimation of reflectivity. In this paper, based on the non- Gaussian statistical character of reflectivity, the negentropy of deconvolution output is expressed as non, polynomial functions to build objective function for blind deconvolution under the unit variance constraint of deconvolution output. Then Particle Swarm Optimization is applied to search for optimum deconvolution operator in order to achieve deconvolution of seismic data. Numerical simulations and actual seismic data processing results prove that, in contrast to the conventional deconvolution methods, the proposed method is suitable for both minimum-phase and mixed-phase wavelet, can obtain a good estimation of reflectivity from seismic data, and can broaden the spectrum of seismic data effectively resulting in the improved resolution.

关 键 词：盲反褶积 反褶积算子 负熵 粒子群算法 混合相位子波 

分 类 号：P631[天文地球—地质矿产勘探]