摘要
基于Gaussian滤波函数和Tikhonov滤波函数的联系,选择Gaussian滤波函数作为正则化矩阵,提出了一种改进的病态问题奇异值修正法——Tikhonov-Gaussian法。通过球体重力模型数据的向下延拓仿真实验,验证了改进的奇异值修正法优于标准的Tikhonov修正法。
Based on the relationship between the Gaussian and Tikhonov filter functions,we choose the Gaussian filter function as the regularization matrix,and propose an improved singular values modification method for ill-posed problem,we term it the Tikhonov-Gaussian method.Simulation results from the downward continuation of the model gravity data show that the improved singular values modification method has significant quality advantages in comparison to the standard Tikhonov method.
出处
《武汉大学学报（信息科学版）》
EI
CSCD
北大核心
2015年第10期1349-1353,共5页
Geomatics and Information Science of Wuhan University
基金
国家自然科学基金资助项目(41171351
61302195)~~
关键词
不适定问题
高斯函数
向下延拓
正则化参数
L曲线
ill-posed problem
Gaussian function
downward continuation
regularization parameter
Lcurve method
