三种算法联合迭代反演求取最佳剩余静校正量
|
摘要
估算剩余静校正量可看作是以激发点及接收点剩余静校正量为模型参数的、以叠加能量为目标函数的全局最优值的搜索问题．当地震记录的剩余静校正量大而信噪比较低时，应用相关统计法不能有效地拾取静校正量．本文分别采用了ＣＭＰ道集内道间互相关法、图像边缘提取的Ｌａｐｌａｃｉａｎ算法及ＤＦＰ算法，三种不同算法联合迭代反演可求取最佳的模型参数，使目标函数值达到或接近于全局最优．理论试算及实际资料处理效果表明，这种方法能够解决低信噪比及大静校正量造成的周期跳跃等静校正问题．
In the final analysis, residual statics estimation is the problem how to search the global optimum point of stack power objective function with regard to model parameters of shot and receiver residual static correction values. Methods of statistical correlation fail to correctly pick up static correction values when residual static correction values are large and S/N (signal to noise) ratio of seismic traces is low. In this paper the global optimum point or it's approximate value can be found by applied the optimum model parameters estimated by iterative inversion of three organized algorithms called CMP (common-median-point) intertrace cross-correlation algorithm、Laplacian's image edge extraction algorithm and DFP algorithm (establishedby Davidon, Fletcher, Powell). It's theoretically and practically shown that the inversion method can prevent cycle skips due to large statics and has extensive practiceapplication values.
In the final analysis, residual statics estimation is the problem how to search the global optimum point of stack power objective function with regard to model parameters of shot and receiver residual static correction values. Methods of statistical correlation fail to correctly pick up static correction values when residual static correction values are large and S/N (signal to noise) ratio of seismic traces is low. In this paper the global optimum point or it's approximate value can be found by applied the optimum model parameters estimated by iterative inversion of three organized algorithms called CMP (common-median-point) intertrace cross-correlation algorithm、Laplacian's image edge extraction algorithm and DFP algorithm (establishedby Davidon, Fletcher, Powell). It's theoretically and practically shown that the inversion method can prevent cycle skips due to large statics and has extensive practiceapplication values.
引文
［1］Hileman，J．A．，Embree，P．，Pflueger．J，C．，Automatedstaticcorrections，CeophysicalProspecting，16，326－358，1968．［2］Wiggins，R．A．，Lamer，K．，Wisecup，D．，Residualstaticsanalysisasagenerallinearinverseproblem，Geophysics，41，922－938，1976．［3］Ronen，S．，Claerbout，J．F．，Residualstaticscstimationbystackpowermaximization，Gcophysics，50，2759─2767，1985．［4］Rothman，D．H．，Automaticestimationoflargeresidualstaticscorrection，Geophysics，51，332─346，1986．［5］徐建华，图像处理与分析，北京：科学出版社，45－97，1992．［6］薛嘉庆，最优化原理与方法，北京：冶金工业出版社，66－132，1986．
版权所有：© 2023 中国地质图书馆 中国地质调查局地学文献中心