多项式与LS-SVM自适应融合的GPS高程拟合方法  

GPS Height Fitting Method Based on Polynomial and LS-SVM Adaptive Fusion

作　　者：李润芝 LI Runzhi(Huizhou Dayawan Economic and Technological Development Zone Surveying and Mapping Product Quality Testing Center,Huizhou 516081,China)

机构地区：[1]惠州大亚湾经济技术开发区测绘产品质量检测中心,广东惠州516081

出　　处：《测绘与空间地理信息》2024年第2期126-128,共3页Geomatics & Spatial Information Technology

摘　　要：针对多项式模型在线状GPS拟合工程存在精度较低问题,本文构建二次曲面拟合与最小二乘支持向量机自适应融合GPS高程异常拟合方法。首先,分别利用多项式与最小二乘支持向量机进行线状工程GPS高程拟合;然后,利用最优加权组合方法将两种拟合模型自适应融合生成新的高程异常值。结果表明:该方法拟合精度明显优于多项式拟合方法及最小二乘支持向量机拟合方法。In view of the low accuracy of polynomial model in linear GPS fitting engineering,a GPS height anomaly fitting method based on adaptive fusion of quadratic surface fitting and least square support vector machine is constructed in this paper.Firstly,poly-nomial and least square support vector machine are used to fit GPS elevation of linear engineering;Then,the optimal weighted combi-nation method is used to adaptively fuse the two fitting models to generate new height anomaly.The results show that the fitting accura-cy of this method is obviously better than polynomial fitting method and least square support vector machine fitting method.

关 键 词：GPS高程 高程异常 多项式拟合 最小二乘支持向量机拟合 

分 类 号：P228.4[天文地球—大地测量学与测量工程]