最小二乘支持向量机在GPS高程转换中的应用

GPS高程是基于WGS84椭球的大地高，而实际工程中所需的高程一般是基于似大地水准面的正常高，两者相差一高程异常。本文结合某GPS控制网控制点的高程资料，利用最小二乘支持向量机（LS—SVM）结合径向基核函数（RBF）对高程异常进行拟合，从而实现GPS高程向正常高的转换。与二次曲面拟合法的转换结果进行比较可知LS--SVM法具有更高的转换精度。

Least Square Support Vector Machine in GPS Height Conversion

GPS height is typically geodetic height with respect to WGS84 ellipsoid, but the normal height （Hnor）is the height above the quasi--geoid, so the difference between the two height systems is called height anomaly. In this paper, some identical survey points of a GPS control net is adopted as the date set in analysis, and then Least Square Support Vector Machine （LS--SVM） with radial basis function （RBF） is proposed to fit the height anomaly so as to convert GPS height to normal height. The results show that the LS--SVR is more promising than conicoid fitting method in converting GPS height.