【摘要】 从考虑地形改正的基础出发,运用神经网络进行GPS水准高程拟合。从节约计算时间的角度提出了新的思路,认为不必深究于过密的DEM模型数据,DEM模型数据只是中间过程,能达到精度要求即可,如果过于强调其精度,相反还会带入DEM高程模型算法本身的系统误差。并提出了DEM模型的密度是否符合拟合要求的初步判定方法。还对神经网络的基本理论作了简单的概括和说明。更多还原

【Abstract】 Based on terrain,the theory of artificial neural network is used in this paper for GPS height collocation.A new theory is given to save the calculation time.This paper points out that excessive dense of DEM model is not necessary as a transition.If DEM model is paid more attention and much dense of DEM model is achieved,the system errors can be bigger and bigger from DEM calculation.Finally,a simple judgment is put forward to judge whether the density of DEM model is enough or not.更多还原