GNSS高程转换拟合方法对比研究  

Comparative study on fitting methods of quasi-geoid

作　　者：王之豪 曾怀恩 张群[4] 王俊杰 李嗣炀 Wang Zhihao;Zeng Huaien;Zhang Qun;Wang Junjie;Li Siyang(National Field Observation and Research Station of Landslides in the Three Gorges Reservoir Area of Yangtze River,China Three Gorges University,Yichang 443002,China;College of Civil Engineering&Architecture,China Three Gorges University,Yichang 443002,China;Hubei Key Laboratory of Construction and Management in Hydropower Engineering,China Three Gorges University,Yichang 443002,China;Natural Resources and Planning Bureau of Yiling District,Yichang 443100,China)

机构地区：[1]湖北长江三峡滑坡国家野外科学观测研究站,湖北宜昌443002 [2]三峡大学土木与建筑学院,湖北宜昌443002 [3]湖北省水电工程施工与管理重点实验室(三峡大学),湖北宜昌443002 [4]宜昌市夷陵区自然资源和规划局,湖北宜昌443100

出　　处：《工程勘察》2025年第5期61-66,共6页Geotechnical Investigation & Surveying

基　　金：国家自然科学基金资助项目(42074005);湖北长江三峡滑坡国家野外科学观测研究站开放基金项目(2018KTL14);湖北省水电工程施工与管理重点实验室(三峡大学)开放基金项目(2023KSD11)。

摘　　要：针对GNSS与水准数据高程转换方法的适用性问题,采用多项式拟合法、SVM拟合法、LSSVM拟合法、Shepard拟合法、BP神经网络拟合法、LSSVM-Shepard拟合法及BP神经网络-Shepard拟合法进行对比研究,以外符合精度与极差评价拟合效果。结果表明,组合模型法在高程转换精度上优于单一模型;各模型在复杂山地区域的拟合精度与稳定性上存在不足,有待改进;针对不同的地形区域,需要具体研究,采用合适的拟合方法提高拟合精度。To address the applicability issues of methods for converting GNSS and leveling data to elevation,a comparative study was conducted using polynomial fitting,SVM fitting,LSSVM fitting,Shepard fitting,BP neural network fitting,LSSVM-Shepard fitting,and BP neural network-Shepard fitting.Except for external compliance accuracy and range,the fitting effects of various models were evaluated.The results indicate that the combined model approach outperforms the single model in terms of height conversion accuracy.However,the fitting accuracy and stability of each model in complex mountainous regions are insufficient and require improvement.Specific research is needed for different terrain areas,and appropriate fitting methods should be adopted to improve fitting accuracy.

关 键 词：高程异常 高程拟合 组合模型 

分 类 号：P228[天文地球—大地测量学与测量工程]