区域内插点的分布对GPS高程异常的影响  

Influence of Interpolation Points on Height Anomaly

作　　者：王军 李仲勤 WANG Jun;LI Zhongqin(Faculty of Geomatics,Lanzhou Jiaotong University,Lanzhou 730070,China;Gansn Provincial Engineering Laboratory for National Geographic State Monitoring,Lanzhon 730070,China)

机构地区：[1]兰州交通大学测绘与地理信息学院,甘肃兰州730070 [2]甘肃省地理国情监测工程实验室,甘肃兰州730070

出　　处：《测绘与空间地理信息》2018年第7期160-163,167,共5页Geomatics & Spatial Information Technology

摘　　要：在GPS高程异常拟合中,选取分布不同的内插点,采用二次曲面、BP神经网络和RBF神经网络3种方法,对两种点位分布情况进行高程异常值的拟合。拟合后的数据采用了格拉布斯准则检验各种拟合模型的外推点数据的粗差情况,并查找剔除粗差。分别对两种情况进行精度分析,评定不同内插点的分布对高程异常的影响。实例分析证明,在实验地区均匀分布的内插点,无论是内插点的精度,还是外推点的推估精度都有明显的提高,合理地选取内插点,有利于高程异常精度的提高。In the fitting of GPS height anomaly,the distribution of different interpolation points,quadratic surface model,BP neural network and RBF neural network are used to fit the height anomaly value of the two point distribution. After fitting the data,we use the criterion to test the gross error of the extrapolated data of different fitting models,and find out the difference. The accuracy analysis of the two cases was carried out to evaluate the influence of different interpolation points on the height anomaly. The example analysis shows that the interpolation points,which are uniformly distributed in the experimental area,have obvious improvement both in the accuracy of the interpolation points and the accuracy of the extrapolation points. The reasonable selection of interpolation points is helpful to improve the accuracy of height anomaly.

关 键 词：GPS高程异常 内插点 二次曲面模型 BP神经网络 RBF神经网络 格拉布斯准则 

分 类 号：P228.4[天文地球—大地测量学与测量工程]