粗差和系统误差显著性检验统计量统一推导新方法  

A New Unified Deriving Method for Gross Error and Systematic Error Significance Test Statistics

作　　者：刘长建[1,2] 王敏 冯绪[1,2] 张熙 LIU Changjian;WANG Min;FENG Xu;ZHANG Xi(Information Engineering University,Zhengzhou 450001,China;Key Laboratory of Smart Earth,Beijing 100080,China;61363 Troops,Xi'an,710054 China)

机构地区：[1]信息工程大学,河南郑州450001 [2]智慧地球重点实验室,北京100080 [3]61363部队,陕西西安710054

出　　处：《测绘科学技术学报》2024年第5期441-445,共5页Journal of Geomatics Science and Technology

基　　金：国家自然科学基金项目(42304043,41374041);智慧地球重点实验室基金项目(KF2023YB01-11)。

摘　　要：基于均值平移模型的数据探测法和附加系统参数的平差法是目前处理粗差和系统误差的重要方法,其函数模型形式一致,故关于二者的显著性检验统计量理论上可做统一推导。不同于已有文献,文中提出一种统一推导的新方法。新方法的思路更符合数理统计中构造显著性检验统计量的一般思路,具有简洁且易于理解的特点。此外,新方法还具有较强的普适性,除参数平差外,需要时还可用于其他平差方法的粗差与系统误差显著性检验统计量推导。The data snooping based on mean shift model and the adjustment method with additional system parameters are currently important methods for handling gross errors and systematic errors.Their function models have the same form,so the significance test statistics can be theoretically derived uniformly.Unlike existing literatures,a new unified deriving method has been proposed.The new method conforms to the general idea of constructing significance test statistics in mathematical statistics,and has the characteristics of simplicity and easy understanding,enriching existing deriving methods.In addition,the new method also has wider applications,and it can also be used to derive gross error and systematic error significance test statistics for other adjustment methods besides parameter adjustment when needed.

关 键 词：粗差 系统误差 显著性检验 统计量 统一推导 新方法 

分 类 号：P207[天文地球—测绘科学与技术]