基于EGM2008的最小二乘支持向量机在高原山区高程拟合中的应用

Application of Least Squares Support Vector Machine Based on EGM2008 in Plateau and Mountainous Areas Elevation Fitting

针对最小二乘支持向量机(LSSVM)单一模型在高原山区高程拟合中精度不高的问题,本文提出了一种基于EGM2008重力场模型的LSSVM"移去恢复"组合方法,以云南某山区的GNSS高程数据为研究对象,通过调整拟合点数目进行高程拟合对比分析。研究结果表明,该组合模型拟合精度高于LSSVM,是一种可行的方法。

Aiming at the problem of the low accuracy of the least square support vector machine(LSSVM)in the elevation fitting of plateau and mountainous areas,the paper proposes a combined method of"removing and restoring"LSSVM based on the EGM2008 gravity field model,uses GNSS elevation data in Yunnan province as the research object,and the fitting comparative analysis is carried out by adjusting the number of elevation fitting points.The results show that the fitting accuracy of the combined model is higher than that of LSSVM,which is a feasible method.