摘要
在反射系数白噪、子波最小相位的假设下,研究基于线性反演的地震反射系数和子波同时估计问题。在Cauchy准则稀疏反演求解中,应用预条件共轭梯度法实现反射系数和子波同时迭代反演。在迭代求解正则化方程时,用共轭梯度法求解相应的原问题,初猜子波求解也使用该策略。模型数据试算与比较,表明了该算法正确而有效。用实际数据检验算法的实用性,经研究表明,预条件共轭梯度法计算的反射系数和子波,要比直接稀疏反演精度高,而且收敛较快,数值稳定,实用性强。
Assuming white noise reflectivity and minimum phase wavelet,the estimation of them simultaneously by linear inversion algorithm is studied.Under the Cauchy criteria,estimating reflectivity and wavelet simultaneously is implemented by the Pre-conditional Conjugate(PCG) schemes.In solving the inversion problems by CG,it is better to solve the original linear system rather than the normalized system for the good matrix condition number.The same measure is taken in solving the initially guessed wavelet.The model data example shows the validation of the algorithm.The advantages of PCG method is validated by real seismic data too.It can be concluded that the PCG sparseness inversion is better than direct sparseness inversion in numerical stability,converge and precision.
出处
《物探化探计算技术》
CAS
CSCD
2006年第3期211-215,共5页
Computing Techniques For Geophysical and Geochemical Exploration
基金
中国科学院知识创新重大项目资助(KZCX1-SW-18)
关键词
地震盲反褶积
Cauchy准则
迭代反演
预条件共轭梯度
seismic blind de-convolution
cauchy criteria
iterating inversion
pre-conditional conjugate gradient
