摘要
以提高GPS高程异常拟合的精度为目标,针对实际工程数据,对BP网络模型进行详细的设计,应用BP神经网络方法进行粗差的剔除和高程异常拟合实验及模型精度的评定,得到较满意的结果。通过与多面函数法得到的结果进行比较,证实该模型可使拟合精度有较大提高。
This paper aims at improving GPS elevation abnormal fitting precision.For the actual engineering data it makes a detailed design for BP neural network model,and it removes the gross error and dose elevation abnoraml fitting experiments using BP nerual network method.Then it makes an assesment on the precision of the model.The results are satisfying.The comparison with the results of Multi-faceted function method proves that the model in the paper improves the fitting precision a lot.
出处
《测绘工程》
CSCD
2010年第4期12-15,共4页
Engineering of Surveying and Mapping
关键词
GPS
高程异常
拟合
BP神经网络
Global Positioning System
elevation abnormal
fitting
back propagation neural network
