LSQR法在位场反演中的分析与评价

梁生贤, 王桥, 焦彦杰, 廖国忠, 郭境

Analysis and evaluation of the potential field inversion using LSQR method

Sheng-Xian LIANG, Qiao Wang, Yan-Jie JIAO, Guo-Zhong LIAO, Jing GUO

表2 合成模型Tikhonov和LSQR正则化求解结果

Table 2 The ruslts of synthetic models for Tikhonov and LSQR regularization

模型 噪声 水平 Tikhonov LSQR λ rλ Aeλ Rms k rk Aek Rms 3% 20.3 0.0609 3.7347 1.43% 9 0.0948 3.7271 2.79% 单一模型 5% 67.3 0.1260 3.7097 3.02% 7 0.1882 3.6964 4.94% 10% 223.6 0.3482 3.7338 7.10% 6 0.3992 3.7210 9.32% 3% 96.5 1.0956 25.5164 2.49% 12 1.2433 25.5316 3.27% 组合模型 5% 223.6 2.0720 25.5418 4.62% 12 2.1079 25.5031 7.20% 10% 518.2 4.2830 25.6148 14.59% 8 4.3580 25.6336 10.46%