中文摘要：

基于柯尔-柯尔频散理论，采用中间梯度上椭球体的复电阻率计算相应的幅频率，讨论椭球体的充电率、中心埋深和长轴倾角对幅频率异常和曲线形态的影响；研究幅频率资料的反演算法。研究结果表明：由于中间梯度剖面数据量非常有限，选择最小二乘迭代拟合算法是合适的；偏导数矩阵采用差分的方法求得，对模型参数进行无量纲处理，提高了反演效率；该反演算法迭代速度快，能够稳定收敛，反演效果较好。

英文摘要：

Based on the theory of Core-Core dispersion, the apparent amplitude frequency was computed using the complex resistivity on elliptical sphere with central gradient array. And some important factors were discussed, such as the charge-ability, the depth of center and the obliquity of long axis, which influence the apparent amplitude frequency greatly or make the form of curve distort. The inversion method of apparent amplitude frequency data was introduced. Jacobian matrix and the coefficient of linear equation were computed by the finite difference method. The results show that the least-square iterative algorithm is optimum for the apparent amplitude frequency data with the central gradient array because the profile data are limited. The model parameters normalized are taken out of their units, which can be used to quicken the inversion velocity. This inversion algorithm has some advantages, such as quick inversion speed, steady convergence and so on.