TSVD截断新方法及其在PolInSAR植被高反演中的应用

发布时间：2018-04-01 07:12

  本文选题：TSVD　切入点：截断参数　出处：《中国矿业大学学报》2017年06期

【摘要】：截断奇异值法(TSVD)通过截断参数截掉病态矩阵中较小的奇异值来改善模型病态性的影响,提高模型参数的估计精度.由均方误差的角度分析可知,TSVD通过引入少量偏差,降低方差,来实现均方误差的下降,截断参数则是改善模型参数估值均方误差的关键因素.通过分析截掉奇异值后,TSVD模型参数估计方差与偏差的变化情况,提出了依据引入偏差量小于降低方差量确定截断参数的方法,理论依据更为充分,可靠性与准确性更高.将采用新方法确定截断参数的TSVD应用到测量坐标解算及PolInSAR植被高反演中,验证了新方法的可行性和有效性,相比于GCV法和L曲线法,新方法确定的截断参数有效提高了TSVD的解算质量,提高了坐标解算和植被高参数反演的精度和可靠性.
[Abstract]:The truncated singular value method (TSVD) improves the effect of model ill-condition by truncating the smaller singular value in the ill-conditioned matrix, and improves the estimation accuracy of model parameters. From the angle of mean square error, it can be seen that TSVD reduces the variance by introducing a small amount of deviation. The truncation parameter is the key factor to improve the estimation mean square error of model parameters. The method of determining the truncation parameter by introducing deviation is less than decreasing variance is put forward, and the theoretical basis is more sufficient. The reliability and accuracy of the new method are higher. TSVD, which uses the new method to determine the truncation parameters, is applied to the calculation of the measured coordinates and the inversion of the vegetation height of PolInSAR. The feasibility and validity of the new method are verified, compared with the GCV method and the L-curve method. The truncation parameters determined by the new method can effectively improve the quality of TSVD solution and improve the accuracy and reliability of coordinate calculation and vegetation high parameter inversion.
【作者单位】： 中南大学地球科学与信息物理学院;湖南科技大学地理空间信息技术国家地方联合工程实验室;
【基金】：国家自然科学基金项目(41531068;41474008;41574006;41674012)
【分类号】：Q948;TN957.52
，

本文编号：1694490