联合QL近似与迭代Born近似法电导率散射成像

对于多维电磁响应的数值模拟和成像反演，积分方程法只需对异常体进行剖分和积分，不涉及复杂的吸收边界条件等问题，具有计算速度快和计算输度高等特点。用于电磁场数值模拟的近似法同样可以用于快速成像反演，它们的模型都是建立在散射电磁场积分方程的线性化基础之上的。联合运用QL近似法与迭代Born近似法进行电导率成像反演，将QL近似法得到的结果作为Born迭代法的初始模型，不用人为的设置初值，尽管方法简单，但却较为有效。

CONDUCTIVITY IMAGING BASED ON COMBINING QL APPROXIMATION WITH ITERATIVE BORN APPROXIMATION

The integral equation method used in the calculation of electromagnetic field requires to calculate the electric field only in small anomalous regions rather than throughout the earth. It doesn't deal with the complex absorption boundary terms, and thus has advantages of quickly calculating speed and high calculation precision in multi dimension electromagnetic inversion. Based on the linearization approximation of electromagnetic integral equation, a few fast EM inversion methods have been developed, which include Born approximation inversion and QL(quasi linear) approximation inversion method. In this paper, the authors develop a new fast conductivity imaging method by combining QL approximation inversion with an iterative Born approximation. In this method, the authors first do QL inversion and obtain a parameter model. Then, the authors do iterative Born inversion using the parameter model obtained in QL inversion as the initial model of the method. Numeric tests show that the new method can give more precision inversion images.