GIS中属性数据不确定性的表示及传播

不确定性是指关于空间过程和特征不能被准确确定的过程，它是自然界各种空间现象自身固有的属性，如何在GIS数据库中合理的表示不确定性一直是学术界关注的疸。根据数据性质的不同，GIS中数据不确定性可分为定位不确定性和属性不确定性。由于属性数据有定量和定性之分，故相应的属性不确定性可以有2种表示方法：一种是用点误差分析的方法，另一种则是用离散型的数据误差的方法来进行评价，因此可以用标准差来表示定量数据的不确定性，用概率矢量和概率面来分别表示矢量和栅格数据模型中的定位数据不确定性。以此为基础可以很快地导出基于确定的算术运算和逻辑叠加等空间分析操作后的误差传播关系。

The expression and propagation of attribute data uncertainty in GIS

Uncertainty, the process that cannot be accurately verified about spatial process and character, is an intrinsic specific property of each natural phenomenon in nature. How describe it in the GIS database reasonably is always concerned. Data uncertainty in GIS can be classified position uncertainty and attribute uncertainty according to the difference of data character. Attribute uncertainty can be expressed by analysis of point error and evaluated of dispersed error because attributes include fixed quantity data and fixed qualitative data. So square error can express the uncertainty in fixed quantity and probability vector and probability surface can express the uncertainty in fixed qualitative data of vector and grid data model respectively. The relation of the propagation error after spatial analysis such as arithmetic operation and logic overlay could be attained easily at the foundation of above.