不同小波基函数在重力梯度异常正演计算中的应用研究

A study of different wavelet basis functions in forward modeling of gravity gradient anomaly

在复杂地质体的重力梯度异常正演计算中,会生成巨型稠密的灵敏度矩阵,相对应的内存存储需求和矩阵向量乘法的耗时都会增加。笔者基于小波变换理论,首先从理论上给出了小波域中的重力梯度异常正演方程,之后选用不同小波基函数进行模型正演对比实验探讨计算效率。研究表明Db3小波基函数能有效减少正演计算中灵敏度矩阵对内存的需求,同时减少矩阵向量乘法操作数。

In the forward modeling of the gravity gradient anomaly of complex geological body,the sensitivity matrix is generated,which corresponds to the increase of the memory storage requirement and the time consuming of matrix vector multiplication. On the basis of the wavelet transform theory,the forward data equations in wavelet domain are given firstly,and then the experimental results are compared with the different wavelet basis functions. The numerical simulation experiments show that the Db3 wavelet basis function can effectively reduce the sensitivity matrix of the forward computation in the memory requirements,while reducing the number of matrix vector multiplication.