摘要
简述了最小二乘配置法的基本原理,并在水准联测点不多的情况下,直接采用平方根函数作为各随机参数间的协方差函数用于推估GPS高程异常值。由于最小二乘配置法的函数模型中同时考虑了非随机变量和随机变量,使得高程异常值的推估精度更高。实例分析也证明了此模型在精度上优于传统的平面拟合模型和协方差推估模型。残差分析表明此方法更适合于同时存在内插和外推高程异常值的情况。
The basic principle of the least squares collocation is briefly introduced.The square root function is used as covariance function for predicting GPS abnormal height with few control points.The precise of GPS abnormal height predicted by least squares collocation is higher than the traditional plane fitting model and covariance prediction model,because the least squares collocation considers both nonrandom variable and random variable.Residua analysis indicates that the least squares collocation is better when there exist interpolation and extrapolation of GPS abnormal height.
出处
《测绘科学》
CSCD
北大核心
2007年第6期102-103,共2页
Science of Surveying and Mapping
基金
国家教育部博士点基金(20040290503):"动态监测数据特征提取与变形预报理论研究"
中国矿业大学科技基金资助(2005B020):"动态变形数据特征分析与建模新理论"
关键词
最小二乘配置
高程异常
协方差函数
推估
least squares collocation
abnormal height
covariance function
prediction
