缺失数据下基于EM算法的GPS高程拟合  

GPS ELEVATION FITTING BASED ON EM ALGORITHM FOR LACK OF DATA

作　　者：林东方[1] 宋迎春[1] 杜琨[1] 肖琴琴[1] 

机构地区：[1]中南大学地球科学与信息物理学院测绘与国土信息工程系,长沙410083

出　　处：《大地测量与地球动力学》2012年第2期139-142,共4页Journal of Geodesy and Geodynamics

基　　金：国家自然科学基金(40874005);教育部博士点基金(200805331086)

摘　　要：GPS高程拟合常因观测条件限制,部分重要拟合点的数据无法获取,致拟合数据缺失或数据量不充足。采用常规方法拟合,会降低似大地水准面的拟合精度。应用EM算法,添加了有益于高程拟合的"潜在数据",即缺失数据的条件期望,有效提高了缺失数据下的GPS高程拟合精度。Restricted by the observing conditions, some important data of GPS elevation fitting missed or not enough. Using conventional methods will reduce the fitting accuracy of the quasi-geoid. By EM algorithm with the addition of ＂potential data＂ , conditional expectation of the missing data, it can improve the accuracy of GPS elevation fitting for lack of data. are often use of the effectively

关 键 词：缺失数据 EM算法 高程拟合 平面相关拟合 二次曲面拟合 

分 类 号：P207[天文地球—测绘科学与技术]