GPS水准多项式拟合自动优选算法

Auto-choosing Algorithm for Polynomial Fitting in GPS-Leveling

GPS水准中,无论采用几何法还是重力法建立高程异常模型,通常都要采用曲面拟合技术。多项式曲面拟合是一种常见的方法,但该法在选取结点、多项式阶次时却存在一定的人为因素。为了克服这一问题,给出了优选算法并进行了计算机编程,实现了对常用502种阶次组合进行自动优选的功能。此外,对该算法确定的拟合正常高精度及可能达到的水准测量等级等用户所关心的问题也进行了探讨。

Surface-fitting method is often adopted to construct height anomaly model in GPS-leveling. In these methods, polynomial fitting method is often used, but there exists some manually choosing factors itself. In order to choose automatically for the best coefficients, an auto-choosing algorithm was given and realized by use of programming. For the 502 common combinations of polynomial surfaces, the method could give automatically the best one. Additionally, some related problems, such as the accuracy of normal height and achievable leveling height level, were studied.