地面核磁共振正则化反演方法研究

不同的模型约束对基于正则化反演技术的地面核磁共振测深数据反演影响很大。利用理论数据讨论了平坦模型和光滑模型约束及与之相关的正则化参数对反演结果的影响规律。模拟表明，无论是高信噪比数据，还是低信噪比数据，反演的含水量在小正则化参数时，对模型约束的依赖性不大，主要受正则化参数影响。大正则化参数时反演分辨率低，而小正则化参数在保证数据拟合精度提高的同时，增加了模型的分辨率，能获得更可靠的含水量分布。

On Regularization Inversion Method of Surface Nuclear Magnetic Resonance Data

To illustrate the influence of different regularization operator on the inversion of water content distribution from surface nuclear magnetic resonance(NMR) signal,synthetic data of noise free and contaminated have been investigated.The water content distributions of two layers have been studied.Numerical results show that for the data with either high or low signal to noise ratio(SNR),inversion solution depends weakly on regularization operators,while strongly upon regularization para-meter.Relatively large regularization parameters cause low resolution of solution,and a small parameter can improve data fitting and enhance the accuracy of inversed water content distribution.