基于监督聚类法的板块格网速度场构建方法研究  

Research on the construction method of plate grid velocity field based on supervised clustering method

作　　者：曲麒霖 秘金钟[2] 徐彦田[2] 王春杨 韩忠倫 李庚辰 QU Qilin;BEI Jinzhong;XU Yantian;WANG Chunyang;HAN Zhonglun;LI Gengchen(College of Geodesy and Geomatics,Shandong University of Science and Technology,Qingdao,Shandong266590,China;Chinese Academy of Surveying&Mapping,Beijing 100036,China;School of Geomatics,Liaoning Technical University,Fuxin,Liaoning 123000,China)

机构地区：[1]山东科技大学测绘与空间信息学院,山东青岛266590 [2]中国测绘科学研究院,北京100036 [3]辽宁工程技术大学测绘与地理科学学院,辽宁阜新123000

出　　处：《测绘科学》2024年第12期20-27,共8页Science of Surveying and Mapping

基　　金：国家重点研发计划课题项目(2021YFB3900802)。

摘　　要：针对复杂板块地区速度场研究成果较少且内容较为陈旧的问题,该文基于内华达大学大地测量实验室全球速度场数据,选择日本及周边区域为例,提出了利用板块分析的监督聚类法对地壳运动复杂区域的水平速度场进行聚类分析,以快速准确地对这些区域进行子块体划分,再利用反距离加权法构建出GPS速度场格网模型,通过此模型精度评估实验结果表明:各板块E、N方向的平均中误差分别为1.1 mm/a、0.88 mm/a,残差绝对值的均值均小于1 mm/a。该方法为复杂板块地区大地测量参考框架动态维持和区域地壳形变分析提供参考。In response to the problem of limited research results and outdated content on the velocity field in complex plate areas,this paper proposes a supervised clustering method using plate analysis based on global velocity field data from the Geodetic Laboratory of the University of Nevada.Japan and surrounding areas are selected as examples to cluster the horizontal velocity field of crustal movement complex areas,in order to quickly and accurately divide these areas into sub blocks.Then,a multi-inverse distance weighted method is used to construct a GPS velocity field grid model.The accuracy evaluation experiment results show that the average mean square error in the E and N directions of each plate is 1.1 mm/a and 0.88 mm/a,respectively,and the mean absolute value of residuals is less than 1 mm/a.This method provides a reference for maintaining the dynamic framework of geodetic measurements and analyzing regional crustal deformation in the surrounding regions of China.

关 键 词：基准站 速度场模型 块体划分 反距离加权法 监督聚类法 格网模型 

分 类 号：P228[天文地球—大地测量学与测量工程]