摘要
地磁基准图制备是实现地磁匹配导航的基础性工作,对提高导航精度具有重大意义。地磁测量技术手段的提高,为地磁数据的获取提供有效途径。文中采用4种方法构建地磁基准图,并对不同算法构建的基准图进行比较和精度评价。首先确定性方法和地统计方法基于数学函数和变异函数反映地磁数据分布的空间结构特性,能在各采样率下取得较优的插值结果;其次低采样率条件下压缩感知算法插值能力最强,实验中均方根误差为14.41 nT,相比泛克里金方法精度提升约38.5%;最后神经网络算法随样本训练数据量增多,精度增加幅度最大,采样率10%~30%情况下,均方根误差从72.85 nT下降至14.38 nT。实验表明,实现高精度地磁基准图制备需依据不同分辨率、不同数据量的地磁数据选用优势更大的插值方法。
The preparation of geomagnetic reference map is the basic work to realize geomagnetic matching navigation,which is of great significance to improve navigation accuracy.The improvement of geomagnetic measurement technology provides an effective way to obtain geomagnetic data.Four kinds of methods are introduced,including traditional deterministic method and geostatistical method of interpolation and new compressive sensing algorithm and neural network algorithm of interpolation.The above four methods are used to construct geomagnetic reference maps respectively,and the comparison and accuracy evaluation of the reference maps constructed by different algorithms are carried out.The result shows that the deterministic method and geostatistical method reflect the spatial structure characteristics of geomagnetic data distribution based on mathematical function and variogram,and can obtain better interpolation results at various sampling rates.Under the condition of low sampling rate,the compressive sensing algorithm has the strongest interpolation ability,and the root mean square error in the experiment is 14.41 nT,which improves the accuracy by about 38.5%compared with the Pan-Kriging method.As the amount of sample training data increases,the accuracy of neural network algorithm increases the most.When the sampling rate is 10%to 30%,the root mean square error decreases from 72.85 nT to 14.38 nT.The results shows that a more advantageous interpolation method should be selected according to the geomagnetic data with different resolution and data volume to achieve the preparation of high precision geomagnetic reference map.
作者
谢心和
赵东明
刘长青
付林威
XIE Xinhe;ZHAO Dongming;LIU Changqing;FU Linwei(Institute of Geospatial Information,Information Engineering University,Zhengzhou 450001,China)
出处
《测绘工程》
2023年第2期13-20,共8页
Engineering of Surveying and Mapping
关键词
地磁匹配导航
地磁基准图构建
插值分析
克里金
压缩感知
神经网络
geomagnetic matching navigation
geomagnetic reference map construction
interpolation analysis
kriging
compressed sensing
the neural network
作者简介
第一作者:谢心和(1998-),男,硕士研究生;通信作者:赵东明(1977-),男,副教授,博士.
