【摘要】 由于最小二乘(LS)残差统计相关,使得粗差发现和定位关系变得相当复杂,且至今未见有效的理论和方法。作者依据在LS平差前进行粗差发现和定位的思路,有意避开LS残差统计相关的难题,从条件平差模型推导出粗差判断方程(GEJE)Vr=JVt+We建立起以上式判断矩阵J为基础的粗差发现和定位的判断准则:1.不能发现含有粗差的观测量;2.能发现含有粗差但不能定位粗差的观测量。　　收稿日期:2002 02 19;修回日期:2002 12 30基金项目:国家自然科学基金资助项目(40271092);香港特别行政区研究基金委员会部分资助项目(香港理工大学5068/99E)作者简介:岑敏仪(1960 ),男,广东顺德人,教授,博士,主要从事误差估计理论、测量数据处理和精密工程测量等研究。531第2期　　　　　　　　　　岑敏仪等:判断观测值粗差能否发现和定位的一种验前方法　　通过算例的检验,证明了判断矩阵J与控制网中观测量数据结构的对应关系,同时说明粗差不能定位的观测量,再好的粗差检测方法也无法正确定位粗差。更多还原

【Abstract】 Owing to the statistical correlation among residuals after adjustment computation by least squares (LS), the relation between gross errors detection and location has become very complex. The effective theory and method of research on the relation are not appeared until now. In this paper, based on the idea that the gross errors detection and location would be conducted before LS adjustment, the authors derive the following gross errors judgement equation (GEJE) from condition equationsVr=JVt+WeBased on judgement matrix J authors gained the following criterions in a surveying system for judging:1. The observations that gross error can’t be detected;2. The observations that gross error can be detected but the gross error can’t be located.The experimental results not only verify that there are a corresponding relation between the judgement matrix J and the data configuration of observation in control network, but also indicate that the best current methods by now are unable to locate gross error correctly for observation in which gross error can be detected but can not be located.更多还原