摘要
GNSS高程拟合常用的是二次曲面拟合法,该方法需要控制点位分布均匀,针对实际作业中受观测条件的影响部分控制点位数据无法获取,影响到GNSS高程测量精度问题,引入期望极大算法(EM算法),提出高斯分布下的EM算法与二次曲面拟合法相结合的组合算法模型,运用高斯分布下的EM算法的二次曲面拟合法对缺失数据的控制点进行建模分析。该组合算法可以获得缺失数据下未知参数的最佳估值,可有效提高水平面的拟合精度。将某区域的高程拟合控制点作为实验数据,结果表明,组合算法模型可以对缺失数据进行高程拟合,检核点最大误差为0.8 cm,组合模型拟合精度较高。
GNSS height fitting is commonly used in the two degree surface fitting method.This method needs uniform distribution of control points.Aiming at the influence of observation conditions in practice,some control points data cannot be obtained,which affects the accuracy of GNSS height measurement.By introducing the expectation maximization algorithm(EM algorithm),a combined algorithm model combining EM algorithm and two degree surface fitting method is proposed.The two degree surface fitting method based on EM algorithm is used to model and analyze the control points of missing data.The combined algorithm can get the best estimation of unknown parameters under missing data,and can effectively improve the fitting accuracy of horizontal plane.The control points of elevation fitting in a certain area are taken as experimental data.The results show that the combined algorithm model can fit the missing data.The maximum error of checkpoint is 0.8 cm,and the fitting accuracy of the combined model is high.
作者
梁安宝
马传宁
周成新
LIANG Anbao;MA Chuanning;ZHOU Chengxin(The First Surveying and Mapping Engineering Institute of Heilongjiang,Harbin 150025,China)
出处
《测绘与空间地理信息》
2020年第1期39-41,44,共4页
Geomatics & Spatial Information Technology
基金
北斗地基增强导航定位与HLJCORS融合服务研究(GA17A301)
黑龙江省对地观测与导航工程技术创新平台(ZY17A02)资助
