Abstract: Spatial distance is an important metric indicator which is commonly used to constrain and represent the relative location relations between spatial objects.Therefore，the representation and computation of spatial distance between spatial objects may have a clear effect on the results of spatial query，spatial reasoning and spatial analysis.Classic Euclidean distance is only suitable for points.Meanwhile，simply extended distance metrics as minimum distance，maximum distance and centroid distance do not take into account the geometric characteristics of spatial objects，such as shape，positional distribution and so on.For this reason，the scholars develop some representative distances for various practical applications，e.g.Hausdorff distance，boundary Hausdorff distance，dual-Hausdorff distance，extended Hausdorff distance，Fréchet distance，turning function distance and area of symmetric difference metric.This paper plays emphasis on the summary of the representation methods of all these distances，pointing out their limitations and adaptabilities，so as to develop more robust distance metrics for application problems in geo-information science.

Key words: spatial distance, spatial analysis, similarity, geographical information system

摘要： 空间目标间的距离是约束和表达空间目标分布关系的一个重要度量指标。如何表达和计算空间目标间的距离将直接影响空间查询、推理和空间分析结果的有效性。经典的欧氏距离只适合于点目标，而简单扩展的最近、最远和质心距离未顾及空间目标的整体形状、位置分布等特征。针对这些传统距离的局限性，学者们基于实际应用问题分别发展了一些较有代表性的距离表达方法，如Hausdorff距离、Hausdorff边界距离、对偶Hausdorff距离、广义Hausdorff距离、Fréchet距离、旋转函数距离以及对称差的面积度量。着重阐述这些距离的表达方法以及适用性，以便于发展更稳健的距离度量方法，更好地解决地理信息科学领域中的实际问题。

关键词: 空间距离, 空间分析, 相似性, 地理信息系统