摘要
如何精确计算大地高与正常高之间的差异即高程异常,一直以来都是测绘领域研究的热点和难点。本文在阐述最小二乘支持向量机模型(LSSVM)原理的基础上,将其应用到高程异常求解过程中,并与传统二次曲面拟合模型就行比较。应用算例结果显示,无论是平坦地区还是山地区域,LSSVM拟合模型在高程拟合中都明显优于二次曲面拟合模型,在GPS高程拟合中,LSSVM拟合模型可达到三、四等甚至有条件的二等几何水准测量精度。
Height anomaly, the difference between the normal height and geodetic height is always a research focus in survey field. This paper described Least Squares Support Vector Machine model （LSSVM） based on the principle, and used it on the solution process of height anomaly, then compared with the traditional quadratic surface model. Experiment showed that for both the fiat areas and mountainous regions, LSSVM model would be significantly better than quadratic surface model. Moreover, in the GPS elevation fit- ting, the accuracy of LSSVM model could be up to three, four, even conditional second geometric leveling accuracy.
出处
《测绘科学》
CSCD
北大核心
2013年第6期166-168,共3页
Science of Surveying and Mapping
