GIS中折线元及面元置信水平的随机模拟  

The Stochastic Simulation of Confidence Levels of Polygon and Area in GIS

作　　者：郭同德[1] 王家耀[2] 魏海平[2] 

机构地区：[1]郑州大学环境与水利学院,河南郑州450052 [2]信息工程大学测绘学院,河南郑州450052

出　　处：《测绘学院学报》2004年第2期136-137,140,共3页Journal of Institute of Surveying and Mapping

摘　　要：置信水平是描述GIS中线元素与面元素的位置不确定性的重要指标之一。文中应用MonteCarlo方法,利用C++中的rand()函数,经过转换,首先产生符合对点误差分布的节点坐标,再按照置信域的构建规则生成大量的随机区域,通过数值判断,算出区域覆盖要素真值的频率,以此频率作为置信水平的近似值,克服了用解析证明方法研究置信水平时结果往往过于保守这一缺陷。本模拟方法有较大的适用范围,为类似的置信水平研究提供了一般方法。The confidence levels of the fundamental factors in GIS are often used to express the uncertainties of these factors. Generally, analysis methods can only the lower bound of the level. Inevitably, the area of the confidence region is larger than needed and the confidence level is higher than the given level based on the analysis method. This paper simulates the confidence level. By employing Monte Carlo method, numerous stochastic regions are produced. By counting the number of the regions that cover the factor, the covering frequency of the regions can be obtained. According to the law of great numbers, the covering frequency is the approximate value of the confidence level. The results given in this paper also show that the Monte Carlo method is convenient in dealing with the confidence level and can be widely used to the similar discussion concerning the confidence level.

关 键 词：随机模拟 置信域 置信水平 

分 类 号：P282[天文地球—地图制图学与地理信息工程]