本文提出了一种等价计算方法,首次将实测位场数据集变换为较小的等价数据集,从而大大节省了CPU时间。与其它传统网格技术相比,等价源数据解释方法是很有竞争力的,因为前者未将位场异常视为调和函数。我们用最小平方迭代算法获得了等价数据集,每次迭代求解一个欠定方程组,对前次迭代选出的所有观测值与最大剩余值进行拟合。通过当前迭代参数计算一组“预测观测值”,并从实测数据中减去这些值,可以计算出剩余值。算法采用Cholesky分解实现,每次得一个修正后的有效解,并处理一个新的基准面。此外,在应用等价层进行插值时,该方法效果最佳,因 In this paper, an equivalent method is proposed. For the first time, the data field of the measured field is transformed into a smaller equivalent data set, which saves CPU time greatly. Compared with other traditional grid technologies, the equivalent source data interpretation method is very competitive because the former does not consider the field anomaly as a harmonic function. We obtained the equivalent dataset using the least-squares iterative algorithm, solved an underdetermined equation set for each iteration, and fitted all observations and maximum residuals selected in the previous iteration. Calculate the residual value by calculating a set of “predicted observations” from the current iteration parameters and subtracting these values ​​from the measured data. The algorithm uses Cholesky decomposition to achieve a modified valid solution and to process a new datum. In addition, this method works best when interpolating with equivalent layers