摘要
为了解决全球导卫星航系统(GNSS)应用于滑坡灾害的地表三维形变监测中,须识别并削弱多路径误差影响的难题,提出一种基于特征分解的非监督分类识别方法:通过对筛选出的特征因子进行特征分解,在得到无关联的正交特征因子后进行聚类处理,以降低特征空间维度数据的冗余性和相关性,进而有效识别并削弱GNSS多路径误差。实验结果表明,所提方法在轮廓系数、聚类结果准确率等指标上均优于常规归一化处理方法;将识别出的受多路径影响较大的观测值剔除后,与原始和归一化处理方案相比,模糊度固定率和水平解定位精度可分别提高10.1%和21.4%,以及6.0%和8.3%。所提方法在多路径误差识别和定位精度提升方面具有一定优势。
In order to address the challenge of identifying and mitigating the effects of multipath errors in the application of global navigation satellite systems(GNSS)for monitoring three-dimensional surface deformation in landslide disasters,the paper proposed an unsupervised classification and recognition method based on feature decomposition:by decomposing the selected feature factors and obtaining unrelated orthogonal feature factors,clustering was performed to reduce the redundancy and correlation of feature space dimension data,effectively identifying and weakening GNSS multipath errors.Experimental results showed that the proposed method would outperform conventional normalization methods in terms of contour coefficient and clustering accuracy;after removing the observed values that were significantly affected by multipath,compared with the original and normalized processing schemes,the ambiguity fixation rate and horizontal solution localization accuracy would be improved by 10.1%and 21.4%,as well as 6.0%and 8.3%,respectively;in short,the proposed method could have certain advantages in multi-path error recognition and positioning accuracy improvement.
作者
卢少骥
舒宝
王利
李新瑞
李东旭
许豪
吴震宇
LU Shaoji;SHU Bao;WANG Li;LI Xinrui;LI Dongxu;XU Hao;WU Zhenyu(School of Geological Engineering and Geomatics,Chang’an University,Xi’an 710054,China;State Key Laboratory of Geographic Information Engineering,Xi’an 710054,China;Key Laboratory of Western China’s Mineral Resources and Geological Engineering,Ministry of Education,Xi'an 710054,China)
出处
《导航定位学报》
北大核心
2025年第4期39-47,共9页
Journal of Navigation and Positioning
基金
国家自然科学基金项目(42127802)
中央高校基本科研业务费专项(300102263202)
陕西省地学大数据与地质灾害防控三秦学者创新团队项目(2022)
陕西省科技创新团队项目(2021TD-51)。
关键词
滑坡
全球卫星导航系统(GNSS)多路径误差
非监督分类
特征分解
归一化
landslides
global navigation satellite system(GNSS)multipath error
unsupervised learning
feature decomposition
norma lization
