点质量融合模型的海空重力数据延拓分析

Extension analysis of sea surface gravity and airborne gravity based on point mass fusion model

针对无高质量的数据对传统的虚拟点质量模型对重力数据改正时,易受系统误差影响及解算不稳定的问题,该文将海空重力数据中的系统误差用非参数分量表达,对点质量模型进行改进,建立了虚拟点质量的半参数模型,引入核估计方法,利用偏残差估计方法,进行三步拟合估计,得到了系统误差和点质量参数的估计值。在分离系统误差的基础上,引入正则化方法改善病态,将建立的模型与正则化方法相结合,建立了综合半参数核估计和正则化方法的点质量融合模型,可在无外部重力时估计系统误差,同时完成多源数据融合。基于EGM2008位模型产生海空重力异常数据,采用线性项和周期项系统误差进行仿真实验,结果表明改进后的点质量模型能够有效利用外部数据,在改善病态和分离系统误差方面提升精度效果明显。并以北卡罗来纳实测海空重力异常数据验证该文方法的有效性,结果表明在无外部数据时,该文方法能有效地分离系统误差,并且精度与引入外部数据改正后的正则化方法相当。

Aiming at the problem that the fusion accuracy is limited by the system error and the solution is unstable when the traditional point mass method has no external data to correct the gravity data,the system error in the sea surface and airborne(sea-air)gravity data was expressed by non-parametric components,and the semi-parametric model of virtual point mass was established to improve the point mass model in this paper.By introducing the kernel estimation method and using the partial residual estimation method,three-step fitting estimation was carried out to obtain the estimated values of system error and point quality parameters.On the basis of separating the system error,the regularization method was introduced to improve the ill-condition.Combining the established model with the regularization method,a point mass fusion model combining semi-parametric kernel estimation and regularization method was established,which could estimate the system error without external gravity and complete multi-source data fusion at the same time.The EGM2008potential model was used to generate sea-air gravity anomaly data,and the linear term and periodic term systematic errors were used for simulation experiments.Experimental results indicated that the improved point mass model could effectively utilize external data,improve the accuracy of ill-condition and separation system errors.At the same time,the effectiveness of the proposed method was verified by using the observed sea-air gravity anomaly data in North Carolina.Experimental results indicated that the proposed method could effectively separate systematic errors without external data,and its accuracy was equivalent to that of the regularization method with external data.