摘要
本文介绍了用神经网络方法转换GPS高程为正常高 ,给出了一种BP神经网络的拓扑结构和算法 ,并与二次多项式曲面拟合方法作了比较分析。经实例验证 ,在较大范围内 ,用神经网络方法转换GPS高程优于二次曲面拟合方法 ,所获得的正常高可满足各种大比例尺测图的精度要求 。
In this paper,a method for conversing GPS height to normal one by means of neural network is proposed.The topological structure and the algorithm of BP neural network are presented and compared with the fitting method which simulates the Geoid(or Quasi Geoid) using Function of manies.Tested by the real case,the conversion of GPS height by the mentioned method is much better than the fitting one.The normal height obtained can satisfy the required precision of various large-scale mapping and could be helpful to engineering practice.
出处
《工程勘察》
CSCD
北大核心
2004年第2期49-51,共3页
Geotechnical Investigation & Surveying
关键词
神经网络
GPS高程
高程异常
高程转换
artifical neural network
GPS height
normal height
height abnormity
height conversion
