大地高转换为正常高的多种算法比较  

Comparison of Various Algorithms for Converting Geodetic Height to Normal Height

作　　者：袁超 YUAN Chao(Jiangsu Institute of Hydrogeology and Marine Geological Exploration,Huai’an 223005,China)

机构地区：[1]江苏省水文地质海洋地质勘查院,江苏淮安223005

出　　处：《测绘与空间地理信息》2023年第3期120-123,127,130,共6页Geomatics & Spatial Information Technology

摘　　要：GPS是一种基于卫星的无线电导航和定位技术,可以为各类用户提供精确的三维坐标,但GPS定位精度仍存在一些不尽如人意的地方。本文从GPS高程测量应用问题着手,介绍了利用平面拟合、二次曲面拟合、最小二乘配置法、神经网络将大地高转换为正常高的算法,并根据这些算法把某地区测量得到的GPS数据进行了拟合处理,从而使实验研究获得了一些技术成果。结果表明,不同阶次的曲面拟合模型拟合效果不同,通常情况下,较高阶次的模型内符合精度更好。GPS is a satellite-based radio navigation and positioning technology,which can provide all kinds of users with accurate three-dimensional coordinates,but GPS positioning accuracy still has some unsatisfactory places.Starting from the application of GPS height measurement,this paper introduces the algorithm of plane fitting,quadric surface fitting,least square collocation and neural network to convert geodetic height to normal height,and according to these algorithms,the GPS data measured in a certain area is fitted,thus making the experimental research obtain some technical achievements.The results show that different order surface fitting models have different fitting effects,and generally,higher order models have better internal fitting accuracy.

关 键 词：GPS高程拟合 高程异常 正常高 最小二乘配置法 神经网络算法 

分 类 号：P228[天文地球—大地测量学与测量工程]