基于EGM2008和BP神经网络的GPS高程转换方法研究

Based on the EGM2008 and BP Neural Network to Research on the Transformation Method of GPS Elevation

当前GPS在平面控制测量中得到了广泛应用,而在高程测量中使用得较少,为了能够使用GPS高程来代替传统的水准测量,就必须进行高程转换。目前,高程转换的方法主要是数学拟合法,而在似大地水准面变化不平稳的地区,这种方法得到的结果却不太理想。本文将EGM2008地球重力场模型与GPS水准相结合起来,运用BP神经网络算法来确定局部似大地水准面,结合具体实例进行拟合试验,并与传统数学拟合法的结果进行比较分析,验证这种方法的优越性与实用性。

At present GPS has been widely used in the horizontal control survey . However, less used in height measurement,in order to be able to use GPS elevation to replace the traditional leveling, it must be converted elevation. At present,the elevation conversion method primarily mathematical fitting, in Quasigeoid change unstable region, the results obtained by this method was less than ideal. This article will EGM20 0 8 Earth gravity lield model with GPS standards combined together,using the BP neural network algorithm to determine the local Quasi geoid,with specific examples to fit test,and legitimate results with traditional mathematics proposed a comparative analysis to verify this advantages and utility of the process.