摘要
由于我国新一代CQG2000似大地水准面不能满足实际需要,本文引入基于径向基的神经网络模型用于似大地水准面的精化,取得了较好的结果。经实例验证:如果拟合点的取样间隔少于5km,参数选取合适,可以得到与四等水准测量精度相当的似大地水准面。
Because China＇s CQG2000 cannot meet actual needs, this paper introduced the RBF neural network model to refine quasi-geoid and achieved good results. It proves that if fitting selection points interval are less than 5 km and parameters are selected properly, it can achieve quasi-geoid which accuracy is equivalent with IVgrade leveling survey.
出处
《电力勘测设计》
2008年第3期20-23,共4页
Electric Power Survey & Design
