Point-Set Topological Spatial Relations 点集拓扑空间关系

原文地址
http://www.dpi.inpe.br/gilberto/references/egenhofer_point_set.pdf

MAX J.EGENHOFER

National Center for Geographic Information and Analysis and Department of Surveying Engineering, 107 Boardman Hall,University of Maine, Orono, Maine 04469, U.S.A.

国家地理信息分析中心与测量工程系，缅因大学博德曼厅107号，美国缅因奥罗诺04469

ROBERT D. FRANZOSA

Department of Mathematics, 417 Neville Hall,University of Maine, Orono, Maine 04469, U.S.A.

美国缅因大学内维尔厅417号数学系，缅因奥罗诺04469

摘要

Abstract. Practical needs in geographic information systems (GIS) have led to the investigation of formal and sound methods of describing spatial relations. Arter an introduction to the basic ideas and notions of topology, a novel theory of topological spatial relations between sets is developed in which the relations are defined in tenns of the intersections of the boundaries and interiors of two sets. By consideringemply and non-emptyas the values of’the intersections, a total of sixteen topological spatial relations is described, each of which can be realized in R2• This set is reduced to nine relations if the sets are restricted to spatial regions, a fairly broad class of subsets of a connected topological space with an application to GIS.

It is shown that these relations correspond to some of the standard set theoretical and topological spatial relations between sets such as equality, disjointness and containment in the interior.

摘要 地理信息系统(GIS)的实际需要导致了对描述空间关系的正式和健全方法的研究。 摘要在介绍拓扑学的基本思想和概念的基础上，提出了一种新的拓扑空间关系理论，该理论将拓扑空间关系定义为两个集合边界的交点和内部的交点。 consideringemply和non-emptyas "十字路口的价值观,共有16个拓扑空间关系描述,每一个都可以实现在R2•这个集合是减少到9如果集限制空间区域的关系,一个相当广泛的一类连接拓扑空间的子集与GIS应用程序。

结果表明，这些关系对应于一些标准集合之间的理论关系和拓扑空间关系，如内部的相等、脱节和包容等。

1. 说明

The work reported here has been motivated by the practical need for a formal understanding ofspatial relations within the realm of geographic information systems (GIS). To display, process or analyze spatial information, users select data from a GIS by asking queries. Almost any GIS query is based on spatial concepts. Many queries explicitly incorporate spatial relations to describe constraints about spatial objects to be analyzed or displayed. For example, a GIS user may ask the following query to obtain information about the potential risks of toxic waste dumps to school children in a specific area: ‘Retrieve all toxic waste dumps which are within 10 miles of an elementary school and located in Penobscot County and its adjacent counties’. The number of elementary schools known to the information system is restricted by using the formulation of constraints. Of particular interest are the spatial constraints expressed by spatial relations such as within 10 miles, in, and adjacent.

这里报告的工作的动机是对地理信息系统(GIS)领域内空间关系的正式理解的实际需要。 为了显示、处理或分析空间信息，用户通过查询从GIS中选择数据。 几乎所有GIS查询都基于空间概念。 许多查询显式地结合空间关系来描述要分析或显示的空间对象的约束。 例如,GIS用户可能会问以下查询获取信息的潜在风险有毒废物倾倒给学校的孩子们在一个特定的面积:检索所有的有毒废物倾倒在10英里的一所小学,位于佩诺布斯科特县及其相邻县”。 利用约束条件的制定来限制信息系统所知的小学数量。 特别令人感兴趣的是空间关系所表达的空间约束，例如10英里内、内和相邻。

The lack of a comprehensive theory of spatial relations has been a major impediment to any GIS implementation. The problem is not only one ofselecting the appropriate terminology for these spatial relations, but also one of determining their semantics. The development of a theory of spatial relations is expected to provide answers to the following questions (Abler 1987):

缺乏全面的空间关系理论是任何地理信息系统实施的主要障碍。 问题不仅在于为这些空间关系选择合适的术语，而且在于确定它们的语义。 空间关系理论的发展有望为下列问题提供答案(Abler 1987):

• What are the fundamental geometric properties of geographic objects needed to describe their relations?

• How can these relations be defined formally in terms offundamental geometric properties?

• What is a minimal set of spatial relations?

• 地理物体的基本几何属性是什么来描述它们之间的关系?

• 如何用基本几何性质来正式定义这些关系?

• 什么是空间关系的最小集合?

In addition to the purely mathematical aspects, cognitive, linguistic and psychological considerations (Talmy 1983, Herskovits 1986) must also be included if a theory about spatial relations, applicable to real-world problems, is to be developed (NCGIA 1989). Within the scope of this paper, only the formal, mathematical concepts which have been partially provided from point-set topology will be considered.

除了纯粹的数学方面，认知、语言和心理方面的考虑(Talmy 1983, Herskovits 1986)也必须包括，如果一个空间关系的理论，适用于现实世界的问题，是发展(NCGIA 1989)。 在本文的范围内，只考虑点集拓扑中部分提供的形式化的数学概念。

The application ofsuch a theory ofspatial relations exceeds the domain ofGIS. Any branch of science and engineering that deals with spatial data will benefit from a formal understanding of spatial relations. In particular, its contribution to spatial logic and spatial reasoning will also be helpful in areas such as surveying engineering, computeraided design/computer-aided manufacturing (CAD/CAM), robotics and very largescale integrated (VLSI) design

这种空间关系理论的应用超出了gis的范围。 任何涉及空间数据的科学和工程分支都将受益于对空间关系的正式理解。 特别是，它对空间逻辑和空间推理的贡献也将有助于测量工程、计算机辅助设计/计算机辅助制造(CAD/CAM)、机器人技术和超大规模集成(VLSI)设计等领域

The variety of spatial relations can be grouped into three different categories: (I) topological relations which are invariant under topological transformations of the reference objects (Egenhofer 1989, Egenhofer and Herring 1990);(2) metric relations in terms of distances and directions (Peuquet and Ci-Xiang 1987); and (3) relations concerning the partial and total order ofspatial objects (Kainz 1990) as described by prepositions such as infront of, behind, above and below (Freeman 1975, Chang et al. 1989, Hernandez 1991). Within the scope of this paper, only topological spatial relations are discussed.

空间关系的多样性可以分为三类:

1. 在参考对象的拓扑变换下不变的拓扑关系(Egenhofer 1989, Egenhofer and Herring 1990);(2)
2. 距离和方向的度量关系(Peuquet and Ci-Xiang 1987); (3)
3. 关于空间对象的偏序和总序的关系(Kainz 1990)，如前面、后面、上面和下面的介词描述(Freeman 1975, Chang et al. 1989, Hernandez 1991)。
   在本文的范围内，只讨论拓扑空间关系。

Formalisms for relations have so far been limited to simple data types in a onedimensional space such as integers, reals, or their combinations, e.g. as intervals (Allen 1983).Spatial data, such as geographic objects or CAD/CAM models, extend in higher dimensions. It has been assumed that a set of primitive relations in such a space is richer, but so far no attempt has been made to explore this assumption systematically.

迄今为止，关系的形式主义仅限于一维空间中的简单数据类型，如整数、实数或它们的组合，如区间(Allen 1983)。 空间数据，如地理对象或CAD/CAM模型，可以扩展到更高的维度。 人们认为，在这样一个空间中，一组原始关系更为丰富，但迄今为止，还没有人试图系统地探索这一假设。

The goal ofthis paper is two-fold. First, to showthat the description of topological spatial relations in terms of topologically invariant properties of point-sets is fairly simple. As a consequence, the topological spatial relation between two point-sets may be determined with little computational effort. Second, to show that there exists a framework within which any topological spatial relation falls. This does not state that the set of relations determined by this -formalism is complete, i.e. humans may distinguish additional relations, but that the formalism provides a complete coverage. i.e. any such additional relation will be only a specialization of one of the relations described.

本文的目的有两个方面。 首先，证明用点集的拓扑不变性质来描述拓扑空间关系是相当简单的。 因此，两个点集之间的拓扑空间关系可以用很少的计算工作量来确定。 第二，证明存在一个任何拓扑空间关系都落在其中的框架。 这并不是说由这种形式主义所决定的一组关系是完整的，即人类可以区分额外的关系，而是说形式主义提供了一个完整的覆盖。 也就是说，任何这样的附加关系将只是所描述关系之一的专门化。

As the underlying data model, subsets of a topological space were selected. The point-set approach is the most general model for the representation of topological spatial regions. Other approaches to the definition of topological spatial relations using different models, such as intervals (Pullar and Egenhofer 1988),or simplicial complexes (Egenhofer 1989), are generalized by this point-set approach.

作为底层数据模型，选择拓扑空间的子集。 点集方法是拓扑空间区域表示中最通用的模型。 使用不同模型定义拓扑空间关系的其他方法，如区间(Pullar and Egenhofer 1988)或单纯复形(Egenhofer 1989)，都被这种点集方法推广。

This paper is organized as follows. The next section reviews previous approaches to defining topological spatial relations. Section 3 summarizes the relevant concepts of point-set topology and introduces the notions used in the remainder of the paper. Section 4 introduces the definition of topological spatial relations and shows their realization in RZ• Section 5 investigates the existence of the relations between two spatial regions, subsets of a topological space with particular application to geographic data handling. In Section 6, the relations within Rn(n>=2) and R1 are compared.

本文的组织结构如下。 下一节回顾之前定义拓扑空间关系的方法。 第3节总结了点集拓扑的相关概念，并介绍了本文其余部分使用的概念。 第4节介绍了拓扑空间关系的定义，并展示了它们在RZ中的实现。第5节研究了两个空间区域之间关系的存在性，这两个空间区域是拓扑空间的子集，在地理数据处理中具有特殊的应用。 第6节比较了Rn(n>=2)与R1之间的关系。

2. Previous work 前期工作

Various collections of terms for spatial relations can be found in the computer science and geography literature (Freeman 1975, Claire and Guptill 1982, Chang et al.1989, Molenaar 1989). In particular, designs of spatial query languages (Frank 1982,Ingram and Phillips 1987, Smith et al. 1987, Herring et al. 1988, Roussopoulos et al.1988)are a reservoirfor informal notations ofspatial relations with verbal explanations in natural language. A major drawback of these terms is the lack of a formal underpinning, because their definitions are frequently based on other expressions which are not exactly defined, but are assumed to be generally understood.

在计算机科学和地理文献中可以找到各种空间关系术语集合(Freeman 1975, Claire and Guptill 1982, Chang et al.1989, Molenaar 1989)。 特别是空间查询语言(Frank 1982,Ingram和Phillips 1987, Smith et al. 1987, Herring et al.1988, Roussopoulos et al.1988)的设计是空间关系的非正式符号和自然语言的口头解释的集合。 这些术语的一个主要缺点是缺乏正式的基础，因为它们的定义经常基于其他没有确切定义的表达式，但假定这些表达式是一般理解的。

Most formal definitions of spatial relations describe them as the results of binary point-set operations. The subsequent review of these approaches will show their advantages and deficiencies. It will be obvious that none of the previous studies has been performed systematically enough to be used as a means to prove that the relations defined provide a complete coverage for the topological spatial relation between two spatial objects. Some definitions consider only a limited subset of representations of ‘spatial objects’, whereas others apply insufficientconcepts to define the whole range of topological spatial relations.

大多数空间关系的正式定义将它们描述为二进制点集运算的结果。 随后对这些方法的审查将显示它们的优点和缺点。 很明显，之前的研究都不够系统，不足以证明所定义的关系完全覆盖了两个空间对象之间的拓扑空间关系。 一些定义只考虑了“空间对象”表示的有限子集，而另一些定义应用的概念不足以定义整个拓扑空间关系范围。

A formalism using the primitives distance and direction in combination with the logical connectors AND, OR and NOT (Peuquet 1986)willnot be considered here. The assumption that every space has a metric is obviously too restrictive so that this formalism cannot be applied in a purely topological setting.

这里不考虑使用原始距离和方向结合逻辑连接器and、OR和NOT (Peuquet 1986)的形式主义。 每个空间都有一个度量标准的假设显然是有局限性的，因此这种形式不能应用于纯粹的拓扑设置中。

The definitions of relationsin terms of set operations use pure set theory to describe topological relations. For example, the following definitions based on point-sets have been given for equal, not equal, inside, outside and intersects in terms of the set operations =, $\ne$, $\subseteq$ and $\bigcap$ [Giiting 1988):

集合操作术语中的关系定义使用纯集合理论来描述拓扑关系。 例如，对于集合操作=、$\ne$、$\subseteq$和$\bigcap$，给出了基于点集的等、不等、内、外、交的定义[Giiting 1988]:

x = y: = points (x) = points (y)

x$\subseteq$y:= points (x) $\ne$ points (y)

x inside y: = points (x)$\subseteq$ points (y)

x outside y:= points (x)$\bigcap$points (y)=$\varnothing$

x intersects y: = points (x)$\bigcap$points (y) $\ne$ $\varnothing$

The drawback of these definitions is that this set of relations is neither orthogonal nor complete. For instance, equal and inside are both covered by the definition of intersects. In contrast, the model of point-sets perse does not allow the definition ofthose relations that are based on the distinction of particular parts of the point-sets such as the boundary and the interior. For example, the relation intersectsistopologically different from that where common boundary points exist, but no common interior points are encountered.

这些定义的缺点是这组关系既不是正交的，也不是完全的。 例如，等号和内号都包含在交点的定义中。 相反，不同的点集模型不允许定义那些关系，这些关系是基于点集的特定部分的区别，如边界和内部。 例如，关系相交在拓扑上不同于存在共同边界点但没有共同内点的关系。

The point-set approach has been augmented with the consideration of boundary and interior so that overlap and neighbor can be distinguished (Pullar 1988):

考虑到边界和内部，点集方法得到了扩充，以便区分重叠和相邻(Pullar 1988):

x overlay y: = boundary(x) $\bigcap$ boundary (y)$\ne$ $\varnothing$ and interior(x)$\bigcap$ interior(y)$\ne$ $\varnothing$

x neighbor y: = boundary (x)$\bigcap$boundary (y) $\ne$ $\varnothing$ and interior(x)$\bigcap$interior (y) = 0

In a more systematic approach, boundaries and interiors have been identified as the crucial descriptions of polygonal intersections (Wagner 1988). By comparing whether or not boundaries and interiors intersect, four relations have been identified: (1) neighborhood where boundaries intersect, but interiors do not; (2) separation where neither boundaries nor interiors intersect; (3) strict inclusion where the boundaries do not intersect, but the interiors do; and (4) intersection with both boundaries and interiors intersecting. This approach uses a single, coherent method for the description of topological spatial relations, but it is not carried out in all its consequences. For example, no distinction can be made between intersection and equality, because for botIi relations boundaries and interiors intersect.

在一个更系统的方法中，边界和内部被确定为多边形交叉的关键描述(Wagner 1988)。 通过比较边界和内部是否相交，确定了四种关系:

1. 边界相交而内部不相交的邻域;
2. 边界和内部都不相交的分离;
3. 严格包含在边界不相交，但内部相交的地方;
4. 与边界和内部交叉的交集。
   这种方法使用一种单一的、连贯的方法来描述拓扑空间关系，但它并不是在所有的结果中进行。 例如，交集和相等之间不能有区别，因为对于botIi关系，边界和内部相交。

3. Point-set topology 点集拓扑学

This model of topological spatial relations is based on the point-set topological notions of interior and boundary. In this section the appropriate definitions and results from point-set topology are presented. Some of the results are stated without proofs. Those proofs are all straightforward consequences ofthe definitions and can befound in most basic topology text books, e.g. Munkres (1966) and Spanier (1966).

该拓扑空间关系模型基于内点集拓扑概念和边界点集拓扑概念。 在本节中，将给出点集拓扑的适当定义和结果。 有些结果是在没有证据的情况下提出的。 这些证明都是定义的直接结果，可以在大多数基本的拓扑教科书中找到，例如Munkres(1966)和Spanier(1966)。

Let X bea set. A topology on X is a collection d ofsubsets of X that satisfies the three conditions: (1) the empty set and X are in d; (2) d is closed under arbitrary unions; and (3) d is closed under finite intersections. A topological space is a set X with a topology d on X. The sets in a topology on X are called open sets, and their complements in X are called closed sets. The collection of closed sets: (1) contains the empty set and X; (2) is closed under arbitrary intersections; and (3)is closed under finite unions.

设X为集合。 X上的拓扑是X子集的集合d，满足以下三个条件:

1. 空集和X在d;
2. d在任意并集下是封闭的;
3. d在有限交点下是封闭的。
   拓扑空间是一个在X上具有拓扑d的集合X。在X上的拓扑中的集合称为开集，它们在X上的补集称为闭集。
   闭集集合:
4. 包含空集和X;
5. 在任意交叉口下封闭;
6. 在有限并集下是封闭的。

Via the open sets in a topology on a set X, a set-theoretic notion of closeness is established. If U is an open set and x$\in$U, then U is said to be a neighborhood of x, This set-theoretic notion of closeness generalizes the metric notion of closeness. A metric d on a set X induces a topology on X, called the metric topology defined by d. This topology issuch that U $\in$ X is an open set if, for each x$\in$U, there is an $\epsilon$ > 0 such that the d-ball of radius earound x is contained in U. A d-ball is the set of points whose distance from x in the metric d is less than s, i.e. {y$\in$X|d(x,y)<$\epsilon$}.

通过对一个集X上的拓扑的开放集，建立了一个集论上的近距离概念。 如果U是一个开放集，xeU，则U是x的一个邻域。这个集论的密度概念推广了密度度量的概念。 指标d在一组X诱发拓扑X,叫做度量拓扑定义为d。这种拓扑issuch U c。X是一个开放的设置,如果每个xeU有e > 0的d-ball半径earound X是包含在美国d-ball X点集的距离度量d小于年代, 即： {y$\in$X|d(x,y)

For the remainder of this paper let X be a set with a topology $\delta$. If S is a subset of X then S inherits a topology from $\delta$.This topology is called the subspace topology and is defined such that U $\subset$ S is open in the subspace topology if, and only if U = $\bigcap$ V for some set V$\in$$\delta$. Under such circumstances, S is called a subspace of X.

在本文的剩余部分，设X是一个拓扑为 $\delta$ 的集合。如果S是X的一个子集，那么S从 $\delta$ 继承了一个拓扑。 这种拓扑称为子空间拓扑，定义为当且仅当某集合V$\in$$\delta$的U = S $\bigcap$ V时，U $\subset$ S在子空间拓扑中是开放的。 在这种情况下，S称为X的一个子空间。

3.1. Interior 内部的

Given Y $\subset$ X, the interior of Y, denoted by Z, is defined to be the union of all open sets that are contained in Y, i.e. the interior of Yis the largest open set contained in Y. y is in the interior of Y if and only if there is a neighborhood of y contained in Y$\in$$Y^0$ if,and only if,there is an open set U such that y $\in$ U $\subset$ Y. The interior of a set could be empty, e.g. the interior ofthe empty set is empty. The interior of X is X itself. If U is open then $U^0$=U. If Z $\subset$ Y then $Z^0$ $\subset$ $Y^0$.

在Y $\subset$ X的条件下，定义Y的内部(用$Y^0$表示)为Y中包含的所有开集的并集，即Y的内部是Y中包含的最大的开集。y的室内y当且仅当存在一个社区中包含y $\in$$\delta$,如果且仅当,有一个开集U, y $\in$ U $\subset$ Y内部的一组可能是空的,如空集的内部是空的。 X的内部是X本身。 如果U是开口的，那么$U^0$=U。 如果Z $\subset$ Y，则$Z^0$ $\subset$ $Y^0$。

3.2. Closure 闭合的

The closure of Y, denoted by $\bar{Y}$, is defined to be the intersection of all closed sets that contain Y, i.e. the closure of Y is the smallest closed set containing Y. It follows that y is in the closure’ of Y if and only if every neighborhood of y intersects Y, i.e.y<Z if and only if U + Y = 0 for every open set U containing y. The empty set is the only set with empty closure. The closure of X is X itself. If C is closed then $\bar{C}$=C. If Z $\subset$ Y then Z $\subset$ Y.

Y的闭包，用 $\bar{Y}$ 表示，定义为包含Y的所有闭集的交点，即Y的闭集是包含Y的最小的闭集。由此可知，y在y的闭包中当且仅当y的每一个邻域相交于y，即y$in$ $\bar{Y}$当且仅当对于每个包含y的开集U $\bigcap$ y = $\varnothing$。空集合是唯一具有空闭包的集合。 X的闭包是X本身。如果C是封闭的，那么$\bar{C}$=C。如果 Z $\subset$ Y 那么 Z $\subset$ Y.

3.3. Boundary 边界

The boundary of Y, denoted by Z, is the intersection of the closure of Y and the closure of the complement of Y, i.e. $\partial Y$=$\bar{Y}$ $\bigcap$ $\overline{A+B}$. The boundary is a closed set. It follows that y is in the boundary of Y if and only if every neighborhood of y intersects both Y and its complement, i.e. y $\in$ $\bar{Y}$ if and only if U$\bigcap$Y $\ne$ $\varnothing$ and U$\bigcap$ (X-Y) $\ne$ $\varnothing$ for every open set U containing y.The boundary can be empty, e.g. the boundaries of both .X and the empty set are empty.

Y的边界用$\partial Y$表示，是Y的闭包与Y的补包的交点，即$\partial Y$=$\bar{Y}$ $\bigcap$ $\overline{A+B}$。 边界是一个封闭的集合。由此得出，y在y的边界上当且仅当y的每一个邻域相交于y及其补集，即y$\in$ $\bar{Y}$当且仅当对于每个包含y的开集U, U$\bigcap$Y $\ne$ $\varnothing$ 且U$\bigcap$ (X-Y) $\ne$ $\varnothing$.边界可以是空的，例如. x和空集合的边界都是空的。

3.4. Relationship between interior, closure and boundary 内部、封闭和边界的关系

The concepts ofinterior, closure and boundary are fundamental to the forthcoming discussions of topological spatial relations between sets. The relationships between interior, closure and boundary are described by the following propositions:

内部、封闭和边界的概念是讨论集合之间拓扑空间关系的基础。 内部、闭包和边界之间的关系由以下命题描述:

Proposition 3.1. $Y^0$ $\bigcap$ $\partial Y$ = $\varnothing$

命题3.1 $Y^0$ $\bigcap$ $\partial Y$ = $\varnothing$

Proof: If x $\in$ Y, then every neighborhood U of x intersects X - Y so that U cannot be contained in Y. As no neighborhood U of x is contained in Y it follows that x$\notin$Y and, therefore, $\partial Y$ $\bigcap$ $Y^0$= $\varnothing$.

证明:如果x$\in$Y，那么x的每个邻域U都相交于X - Y，因此U不包含在Y中。由于Y中不包含邻域U，因此x$\notin$Y，因此$\partial Y$ $\bigcap$ $Y^0$= $\varnothing$.

Proposition 3.2. $Y^0$ $\bigcup$ $\partial Y$ = $\bar{Y}$

命题3.2. $Y^0$ $\bigcup$ $\partial Y$ = $\bar{Y}$

Proof: $Y^0$ $\subset$ Y $\subset$ $\bar{Y}$ and, by definition, $\partial Y$ $\subset$ $\bar{Y}$. As $Y^0$ and $\partial Y$ are both subsets of $\bar{Y}$ it follows that ($Y^0$ $\bigcup$ $\partial Y$) $\subset$ $\bar{Y}$. To show that $\bar{Y}$ $\subset$ ( $Y^0$ $\bigcup$ $\partial Y$ ), let x $\in$ $\bar{Y}$ and assume that x $\notin$ $Y^0$ . It is shown that x $\in$ $\partial Y$ which, since x $\in$ $\bar{Y}$, only requires showing that x$\in$ $\overline{A+B}$.x $\notin$ $Y^0$ implies that every neighborhood of x is not contained in $\bar{Y}$ , therefore, every neighborhood of x intersects X - Y, implying that x$\in$(X - Y).So x $\in$ $\partial Y$ . Thus if x $\in$ $\bar{Y}$ and x $\notin$ $Y^0$, then x $\in$ $\partial Y$ ,and it follows that $\bar{Y}$ $\subset$ ($Y^0$ $\bigcup$ $\partial Y$). Thus $Y^0$ = $\partial Y$.

证明:$Y^0$ $\subset$ Y $\subset$ $\bar{Y}$，根据定义，$\partial Y$ $\subset$ $\bar{Y}$。由于$Y^0$和$\partial Y$都是$\bar{Y}$的子集，因此($Y^0$ $\bigcup$ $\partial Y$) $\subset$ $\bar{Y}$。为了证明 $\bar{Y}$ $\subset$ ( $Y^0$ $\bigcup$ $\partial Y$ )，设 x $\in$ $\bar{Y}$ 并假设x $\notin$ $Y^0$ 。 它证明了x $\in$ $\partial Y$，因为x $\in$ $\bar{Y}$，只需要证明x$\in$ $\overline{A+B}$。x $\notin$ $Y^0$表示x的每一个邻域不包含在$\bar{Y}$中，因此x的每一个邻域相交于X - Y，这意味着x$\in$(X - Y) 。所以x $\in$ $\partial Y$。 因此，如果 x $\in$ $\bar{Y}$ 且 x $\notin$ $Y^0$，则x $\in$ $\partial Y$，则$\bar{Y}$ $\subset$ ($Y^0$ $\bigcup$ $\partial Y$)。 因此$Y^0$ = $\partial Y$

3.5 Separation 相离

The concepts of separation and connectedness are crucial for establishing the forthcoming topological spatial relations between sets.Let Y $\subset$ X. A separationof Yis a pair A, B ofsubsets of X satisfying the following three conditions:

1. A $\ne$ $\varnothing$ and B $\ne$ $\varnothing$;
2. A $\cup$ B = Y
3. $\bar{A}$ $\cap$ B = $\varnothing$ and A $\cap$ $\bar{B}$=$\varnothing$

分离和连通性是建立集合间拓扑空间关系的关键概念。 设Y $\subset$ X, Y的A分离是X的子集A, B的一对，满足以下三个条件:

1. A $\ne$ $\varnothing$ and B $\ne$ $\varnothing$;
2. A $\cup$ B = Y
3. $\bar{A}$ $\cap$ B = $\varnothing$ and A $\cap$ $\bar{B}$=$\varnothing$

If there exists a separation of Y, then Y is said to be disconnected, otherwise Y is said to be connected. If Y is the union of two nonempty disjoint open subsets of X, then it follows that Y is disconnected.If C is connected and C $\subset$ D $\subset$ $\bar{C}$, then D is connected. In particular, if C is connected, then $\bar{C}$ is connected; however, $\partial C$ and C${}^{\circ }$ need not be connected.

如果Y分离，则说Y是断开的，否则说Y是连接的。 如果Y是X的两个非空不相交开子集的并集，则可以得出Y是不相交的。如果连接了C并且C $\subset$ D $\subset$ $\bar{C}$，那么D是连接的。 特别地，如果连接了C，则C是连接; 然而，$\partial C$和C${}^{\circ }$不需要连接。

Proposition 3.3. If A, B form a separation of Y and if Z is a connected subset of Y, then either Z $\subset$ A or Z $\subset$ B.

命题3.3。 如果A, B构成Y的一个分离，如果Z是Y的一个连通子集，那么 Z $\subset$ A 或 Z $\subset$ B。

Proof: By assumption, Z is a subset of the union of A and B, i.e. Z $\subset$ a $\cap$ B. It is shown that the intersection between Z and one of A or B is empty, i.e.either Z - B = 0 or Z - A= 0.Suppose not, i.e. assume that both intersections are non-empty. Let C=Z $\cap$ A and D=Z $\cap$ B. Then C and D are both non-empty and C $\cup$ D=Z. As $\bar{C}$ $\subset$ $\bar{A}$, D $\subset$ B, and A $\cup$ B=$\varnothing$ (because A, B is a separation of Y), it follows that $\bar{C}$ $\cup$ D = $\varnothing$. Similarly, C $\cup$ $\bar{D}$= 0; therefore, C and D form a separation of Z, contradicting the assumption that Z is connected. So either Z $\cup$ B= $\varnothing$ or Z $\cup$ A= $\varnothing$, implying that either Z $\subset$ A or Z $\subset$ B.

证明: 假设Z是A与B并集的子集，即Z $\subset$ a $\cap$ B。证明Z与a或B中的一个交为空，即Z $\cap$ B = $\varnothing$ 或Z $\cap$ A = $\varnothing$。假设不是，即假设两个交点都是非空的。设C=Z $\cap$ A, D=Z $\cap$ B，则C和D都是非空的，且C $\cup$ D=Z。 当$\bar{C}$ $\subset$ $\bar{A}$, A $\cup$ B=$\varnothing$(因为A, B是Y的一个分离点)，可以得出$\bar{C}$ $\cup$ D = $\varnothing$。 同理，C $\cup$ $\bar{D}$= 0; 因此，C和D形成了Z的分离，这与Z是连通的假设相矛盾。 要么Z $\cup$ B= $\varnothing$，要么Z $\cup$ A= $\varnothing$，意味着要么Z $\subset$ A，要么Z $\subset$ B。

A subset Z of X is said to separate X if X - Z is disconnected. The following separation result gives simple conditions under which the boundary of a subset of X separates X.

如果X - Z断开连接，则X的一个子集Z被称为分离X。 下面的分离结果给出了X子集的边界分离X的简单条件。

Proposition 3.4. Assume Y $\subset$ X. If Y${}^{\circ }$ $\ne$ $\varnothing$ and Y $\ne$ X, then Y${}^{\circ }$ and X - $\bar{Y}$ form a separation of X -$\partial Y$, and thus $\partial Y$ separates X.

命题3.4。 假设Y $\subset$ X，如果Y${}^{\circ }$ $\ne$ $\varnothing$, Y $\ne$ X，那么Y${}^{\circ }$和X - $\bar{Y}$ 形成X -$\partial Y$的分离，因此$\partial Y$分隔了X。

Proof: By assumption, Y${}^{\circ }$ and X - $\bar{Y}$ are non-empty. Clearly, they are disjoint open sets. Proposition 3.2 implies that X -$\partial Y$= Y${}^{\circ }$+(X - $\bar{Y}$). It follows that Y${}^{\circ }$ and X - $\bar{Y}$ form a separation of X -$\partial Y$.

证明:假设Y${}^{\circ }$和X - $\bar{Y}$ 非空。 显然，它们是不相交的开放集。 命题3.2表明，X -$\partial Y$= Y${}^{\circ }$+(X -$\bar{Y}$)，因此，Y${}^{\circ }$和X -$\bar{Y}$构成了X - $\partial Y$ 的分离。

3.6 Topoiogical equivalence 拓扑等价性

The study of topological equivalence is central to the theory of topology. Two topological spaces are topologically equivalent(homeomorphic or ofthe same topological type) if there is a bijective function between them that yields a bijective correspondence between the open sets in the respective topologies. Such afunction, which is continuous with a continuous inverse,iscalled a homeomorphism. Examples of homeomorphisms are the Euclidean notions of translation, rotation, scale and skew. Properties of topological spaces that are preserved under homeomorphism are called topological invariants of the spaces. For example, the property of connectedness is a topological invariant.

拓扑等价性的研究是拓扑理论的核心。 如果两个拓扑空间之间存在一个双射函数，使其在各自的拓扑上的开集之间产生一个双射对应，则这两个拓扑空间是拓扑等价的(同胚或相同拓扑类型)。 这种具有连续逆的连续函数称为同胚。 同胚的例子有欧几里德的平移、旋转、比例和倾斜的概念。 在同胚下保持的拓扑空间的性质称为该空间的拓扑不变量。 例如，连通性是一个拓扑不变量。

4. A framework for tbe description of topological spatial relatiollS 拓扑空间关系的描述框架

This model describing the topological spatial relations between two subsets, A and .B, of a topological space X is based on a consideration of the four intersections of the boundaries and interiors of the two sets A and B,i.e. $\partial A$ $\cap$ $\partial B$, A${}^{\circ }$ $\cap$ B${}^{\circ }$, $\partial A$ $\cap$ B ${}^{\circ }$ and A ${}^{\circ }$ $\cap$ $\partial B$.

该模型描述了一个拓扑空间X的两个子集A和。B之间的拓扑空间关系，是基于考虑了两个子集A和B的边界和内部的四个交点，即 $\partial A$ $\cap$ $\partial B$, A${}^{\circ }$$ $\cap$ B${}^{\circ }$, $\partial A$ $\cap$ B${}^{\circ }$ and A${}^{\circ }$ $\cap$ $\partial B$。

Definition 4.1. Let A, B be a pair of subsets of a topological space X. A topological spatial relation between A and B is described by a four-tuple of values of topological invariants associated to each of the four sets A=B, respectively

定义4.1。 设A、B是一个拓扑空间X的一对子集。A和B之间的拓扑空间关系由与这四个集合 $\partial A$ $\cap$ $\partial B$, A${}^{\circ }$$ $\cap$ B${}^{\circ }$, $\partial A$ $\cap$ B${}^{\circ }$ and A${}^{\circ }$ $\cap$ $\partial B$ 中的每一个相关联的拓扑不变量值的四元组来描述

A topological spatial relation between two sets is preserved under homeomorphism of the underlying space X. Specifically, if f:A $\to$ B is a homeomorphism and A, B $\subset$ X, then $\partial A$ $\cap$ $\partial B$, A${}^{\circ }$$ $\cap$ B${}^{\circ }$, $\partial A$ $\cap$ B${}^{\circ }$ and A${}^{\circ }$ $\cap$ $\partial B$ are mapped homeomorphically onto $\partial f(A)$ $\cap$ $\partial f(B)$ , f(A)${}^{\circ }$ $\cap$ f(B)${}^{\circ }$,$\partial f(A)$ $\cap$ f(B)${}^{\circ }$,f(A)${}^{\circ }$ $\cap$ $\partial f(B)$, respectively. Since the topological spatial relation is defined in terms of topological invariants of these intersections, it follows that the topological spatial relation between A and B in X is identical to the topological spatial relation between f(A) and f(B) in Y.

在基础空间X的同胚下保持两个集合之间的拓扑空间关系。具体地说，如果f:A $\to$ B是同胚，且A, B $\subset$ X，则 $\partial A$ $\cap$ $\partial B$, A${}^{\circ }$ $\cap$ B${}^{\circ }$, $\partial A$ $\cap$ B${}^{\circ }$ and A${}^{\circ }$ $\cap$ $\partial B$ 分别被同胚映射到$\partial f(A)$ $\cap$ $\partial f(B)$ , f(A)${}^{\circ }$ $\cap$ f(B)${}^{\circ }$,$\partial f(A)$ $\cap$ f(B)${}^{\circ }$,f(A)${}^{\circ }$ $\cap$ $\partial f(B)$上。由于拓扑空间关系是用这些交点的拓扑不变量来定义的，因此可以得出X中A和B的拓扑空间关系与Y中f(A)和f(B)的拓扑空间关系相同。

A topological spatial relation is denoted here by a four-tuple (,,,). The entries correspond in order to the values of topological invariants associated to the four set-intersections. The first intersection is called the boundary-boundary intersection,the second intersection the interior-interior intersection, the third intersection the boundary-interior intersection, and the fourth intersection the interior-boundary intersection.

拓扑空间关系在这里用一个四元组(，，，)来表示。 这些项对应于与四个集合交点相关的拓扑不变量的值。 第一个交点称为边界-边界交点，第二个交点称为内部-内部交点，第三个交点称为边界-内部交点，第四个交点称为内部-边界交点。

4.1 Topological spatial relationsfrom empty/non-empty set-intersections 来自空/非空集交集的拓扑空间关系

As the entries in the four-tuple, properties of sets that are invariant under homeomorphisms are considered. For example, the properties empty and non-empty are set-theoretic, and therefore topologically invariant. Other invariants, not considered in this paper, are the dimension of a set and the number of connected components (Munkres 1966). Empty/non-empty is the simplest and most general invariant so that any other invariant may be considered a more restrictive classifier.

作为四元组中的元素，研究了在同胚下不变集的性质。 例如，空的和非空的属性是集论的，因此是拓扑不变的。 其他不变量在本文中没有考虑，是集合的维数和连通分量的数量(Munkres 1966)。 空/非空是最简单和最一般的不变式，因此任何其他不变式都可以被视为限制性更强的分类器。

For the remainder of this paper, attention is restricted to the binary topological spatial relations defined by assigning the appropriate value of empty ( $\varnothing$ ) and non-empty ( $\neg \varnothing$ ) to the entries in the four-tuple. The 16 possibilities from these combinations are summarized in table 1.

在本文的其余部分中，我们只关注二元拓扑空间关系，它是通过将适当的空值( $\varnothing$ )和非空值( $\neg \varnothing$ )赋给四元组中的项来定义的。 表1总结了这些组合的16种可能性。

A set is either empty or non-empty; therefore, it is clear that these 16 topological spatial relations provide complete coverage, that is, given any pair of sets A and B in X,there is always a topological spatial relation associated with A and B. Furthermore, a set cannot simultaneously be empty and non-empty, from which follows that the 16 topological spatial relations are mutually exclusive, i.e. for any pair of sets A and B in X,exactly one of the 16 topological spatial relations holds true.

一个集合要么为空，要么为非空; 由此可见，这16个拓扑空间关系具有完全复盖性，即给定X中任意一对集合A、B，总存在与A、B相关联的拓扑空间关系，且集合不可能同时为空集和非空集， 由此可知，这16个拓扑空间关系是相互排斥的，即对于X中的任意一对集合A和B，这16个拓扑空间关系中恰好有一个成立。

In general, each of the 16 spatial relations can occur between two sets. Depending on various restrictions on the sets and the underlying topologicalspace, the actual set of existing topological spatial relations may be a subset of the 16 in table 1. For general point-sets in the plane $R^2$, all 16 topological spatial relations can be realized (figure 1).

一般来说，16种空间关系中的每一种都可以发生在两个集合之间。 根据对集合和底层拓扑空间的各种限制，现有拓扑空间关系的实际集合可能是表1中16个拓扑空间关系的子集。 对于平面$R^2$中的一般点集，16种拓扑空间关系均可实现(图1)。

4.2. Influence of the topological space on the relations 拓扑空间对关系的影响

The setting, i.e.the topologicalspace X in which A and B lie, plays an important role in the spatial relation between A and B. For example, in figure 2 (left panel) the two sets A and B have the relation ($\varnothing$,$\neg \varnothing$,$\neg \varnothing$,$\neg \varnothing$) as subsets of the line.The same configuration shows a different relation between the two sets when they are embedded in the plane (figure 2,right panel). Assubsets of the plane, the boundaries of A and B are equal to A and B, respectively, and the interiors are empty, i.e. $\partial A$ = A , A${}^{\circ }$ = $\varnothing$,$\partial B$ = B,B${}^{\circ }$ = $\varnothing$.It follows that in the plane the spatial relation between the two sets A and B is ($\neg \varnothing$,$\varnothing$,$\varnothing$,$\varnothing$).

集合，即A和B所在的拓扑空间X，在A和B的空间关系中起着重要的作用。例如，在图2(左面板)中，A和B两个集合具有($\varnothing$,$\neg \varnothing$,$\neg \varnothing$,$\neg \varnothing$)的关系作为直线的子集。同样的配置显示了两个集合嵌入平面时的不同关系(图2，右面板)。 作为平面的子集，A和B的边界分别等于A和B，内部为空，即 $\partial A$ = A , A${}^{\circ }$ = $\varnothing$,$\partial B$ = B,B${}^{\circ }$ = $\varnothing$。由此得出平面上A和B两个集合的空间关系为($\neg \varnothing$,$\varnothing$,$\varnothing$,$\varnothing$)。

The 16 specifications ofbinarytopological relationsbasedon thecriteriaofemptyand non-empty intersections of boundariesand interiors.
二元拓扑关系的16个规范基于边界和内部的空和非空相交的标准。

Figure 1. Examples ofthe 16 binary topological spatial relations based on the comparison of empty and non-empty set-intersections between boundaries and interiors.
图1所示。 以16种二元拓扑空间关系为例，比较边界与内部之间的空集与非空集相交。

Figure 2. The same configuration of the two sets A and B with (leftpanel) the topological spatial relation ($\varnothing$,$\neg \varnothing$,$\neg \varnothing$,$\neg \varnothing$) when embedded in a line and (right panel) ($\neg \varnothing$,$\varnothing$,$\varnothing$,$\varnothing$) in a plane.

5. Topological relations between spatial regions 空间区域间的拓扑关系

It is the aim of this paper to model topological spatial relations that occur between polygonal areas in the plane; therefore, the topological space X and the sets under consideration in X are restricted. These restrictions are not too specific and the only assumption that is made about the topological space X is that it is connected. This guarantees that the boundary of each set of interest is not empty.

本文的目的是建立平面多边形区域间的拓扑空间关系模型; 因此，拓扑空间X和X中考虑的集合受到限制。 这些限制不是很具体，关于拓扑空间X的唯一假设是它是连通的。 这保证了每一组利益的边界不是空的。

The sets of interest are the spatial regions, defined as follows:
感兴趣的集合是空间区域，定义如下:

Definition 5.1. Let X be a connected topological space. A spatial region in X is a non-empty proper subset A of X satisfying (1) B is connected and (2) A=C

定义5.1。 设X为连通拓扑空间。 X中的空间区域是满足(1)A${}^{\circ }$ 是连通的，(2)A= $\bar{A}$${}^{\circ }$的X的非空固有子集A

It follows from the definition that the interior of each spatial region is non-empty. Furthermore, a spatial region isclosed and connected as it is the closure of a connected set.Figure 3 depicts sets in the plane which, by failing to satisfy either condition (1)or condition (2)in Definition 5.1,are not spatial regions.A and B are not spatial regions,because A${}^{\circ }$ and B${}^{\circ }$ are not connected, respectively. C and D are not spatial regions, because they fail to satisfy condition (2), i.e. C $\ne$ $\overline{C^{\circ }}$ and D $\ne$ $\overline{D^{\circ }}$. The latter sets are needed to realize the topological spatial relations $r_2$, $r_5$, $r_8$,$r_9$, $r_{12}$ and $r_{13}$ in the plane.

根据定义，每个空间区域的内部都是非空的。 此外，空间区域是封闭和连通的，因为它是连通集的闭集。图3描述了平面上的集合，由于不满足定义5.1中的条件(1)或条件(2)，这些集合不是空间区域。A和B不是空间区域，因为 A${}^{\circ }$和B${}^{\circ }$分别没有连接。 C和D不是空间区域，因为它们不满足条件(2)，即C $\ne$ $\overline{C^{\circ }}$和D $\ne$ $\overline{D^{\circ }}$，需要后者集合来实现平面内的拓扑空间关系$r_2$, $r_5$, $r_8$,$r_9$, $r_{12}$ 以及 $r_{13}$。

The following proposition impliesthat the boundary of each spatial region is nonempty.

下面的命题意味着每个空间区域的边界是非空的。

Proposition 5.2. If A is a spatial region in X then $\partial A$ $\ne$ $\varnothing$.

命题5.2。 如果A是X中的一个空间区域，那么 $\partial A$ $\ne$ $\varnothing$.

Proof: $A^{\circ }$ $\ne$ $\varnothing$ A= $\bar{A}$ since A isclosed, and A $\ne$ X by definition ofa spatial region. From proposition 3.4 it follows that $A^{\circ }$ and X -A form a separation of X -$\partial A$. If $\partial A$ =$\varnothing$ then the two sets form a separation of X, which is impossible since X is connected; therefore, $\partial A$ $\ne$ $\varnothing$.

证明:由于A是封闭的，$A^{\circ }$ $\ne$ $\varnothing$ A= $\bar{A}$, A $\ne$ X由空间区域定义。 从命题3.4可以得出，$A^{\circ }$和X -A构成X -$\partial A$的分离。 如果$\partial A$ =$\varnothing$，则两个集合形成X的分离，这是不可能的，因为X是连通的; 因此，$\partial A$ $\ne$ $\varnothing$。

5.1. Existence of region relations 区域关系的存在性

The framework for the spatial relations between point-sets carries over to spatial regions, however, not all of the 16 relations between arbitrary point-sets exist between two spatial regions. From the examples in figure I it is concluded that at least the relations $r_0$, $r_1$, $r_3$,$r_6$, $r_7$ ,$r_{10}$,$r_{11}$,$r_{14}$ and $r_{15}$ exist between two spatial regions. The following proposition shows that these nine topological spatial relations are the only relations that can occur between spatial regions.

点集之间的空间关系框架也适用于空间区域，但是，任意点集之间的16种关系并非全部存在于两个空间区域之间。 从图一的例子可以得出结论，两个空间区域之间至少存在关系$r_0$, $r_1$, $r_3$,$r_6$, $r_7$ ,$r_{10}$,$r_{11}$,$r_{14}$ 和 $r_{15}$。 下面的命题表明，这九种拓扑空间关系是空间区域之间唯一可能出现的关系。

Proposition 5.3. For two spatial regions the spatial relations $r_2$, $r_4$, $r_5$,$r_8$, $r_9$ ,$r_{12}$ and $r_{13}$ cannot occur.

命题5.3。 对于两个空间区域，空间关系$r_2$, $r_4$, $r_5$,$r_8$, $r_9$ ,$r_{12}$ 以及$r_{13}$不能发生。

Proof: This begins by proving that if the boundary-interior or interiorboundary intersection is non-empty then the interior-interior intersection between the same two regions is also non-empty. This implies that the six topological spatial relations $r_4$, $r_5$,$r_8$, $r_9$ ,$r_{12}$ and $r_{13}$ all with empty interior-interior and non-empty boundary-interior or interior-boundary intersections, cannot occur.

证明:首先证明如果边界-内部或内部边界的交点是非空的，那么同一两个区域之间的内部-内部交点也是非空的。 这意味着不可能出现内部空-内部和非空边界-内部或内部-边界交叉点的6个拓扑空间关系$r_4$, $r_5$,$r_8$, $r_9$ ,$r_{12}$ 以及$r_{13}$。

图3. 平面中的点空间区域集合。

Let A and B be spatial regions for which A+B=0. It is shown that A-B=0.Thus if the boundary-interior intersection is non-empty, then the interior-interior intersection is also non-empty. It also follows that if the interior-boundary intersection is non-empty. then the interior-interior intersection is also non-empty.Using proposition 3.2, $A^{\circ }$ $\cup$ $\partial A$ = $\bar{A}$, $A^{\circ }$ + $\partial A^{\circ }$ = $\overline{A^{\circ }}$。 $\bar{A}$ =A=$\overline{A^{\circ }}$, so $A^{\circ }$ $\cup$ $\partial A$=$A^{\circ }$ $\cup$ $\partial A^{\circ }$. Furthermore, by proposition 3.1,A ^\circ$ $\cap$ $\partial A^{\circ }$=$\varnothing$, $A^{\circ }$ $\cap$ $\partial A$=$\varnothing$. It follows that$\partial A^{\circ }$ =$\partial A$. Now let X $\in$ $\partial A$ $\cap$ $B^{\circ }$, then X $\in$ $\partial A^{\circ }$ and since $B^{\circ }$ is open and contains x, it follows that $A^{\circ }$ $\cap$ $B^{\circ }$ $\ne$ $\varnothing$.

设A和B是 $\partial A$ + $B^{\circ }$ $\ne$ $\varnothing$ 的空间区域。 得到 $A^{\circ }$ + $B^{\circ }$ $\ne$ $\varnothing$。利用命题3.2,$A^{\circ }$ $\cup$ $\partial A$ = $\bar{A}$, $A^{\circ }$ + $\partial A^{\circ }$ = $\overline{A^{\circ }}$。 $\bar{A}$ =A=$\overline{A^{\circ }}$，所以 $A^{\circ }$ $\cup$ $\partial A$=$A^{\circ }$ $\cup$ $\partial A^{\circ }$。 更进一步，根据命题3.1, $A^{\circ }$ $\cap$ $\partial A^{\circ }$=$\varnothing$, $A^{\circ }$ $\cap$ $\partial A$=$\varnothing$。 可以得出 $\partial A^{\circ }$ =$\partial A$ 。 现在让X $\in$ $\partial A$ $\cap$ $B^{\circ }$，然后X $\in$ $\partial A^{\circ }$，因为$B^{\circ }$是开放的，包含X，所以$A^{\circ }$ $\cap$ $B^{\circ }$ $\ne$ $\varnothing$。因此，如果边界-内部交点是非空的，那么内部-内部交点也是非空的。 也可以得出，如果内部边界的交点是非空的。 那么内部-内部交叉口也是非空的。

Next it is proved that if the boundary-boundary intersection is empty and the interior-interior intersection is non-empty, then either the boundary-interior or the interior-boundary intersection is non-empty. This impliesthat the spatial relation $r_2$ with a non-empty interior-interior intersection and empty intersectionsfor boundary-boundary, boundary-interior and interior-boundary, cannot occur.This willcomplete the proof of the proposition’.Let A and B be spatial regions such that $\partial A$ $\cap$ $\partial B$ $\ne$ $\varnothing$ and $A^{\circ }$ $\cap$ $B^{\circ }$ $\ne$ $\varnothing$. It is shown that if $\partial A$ $\cap$ $B^{\circ }$= $\varnothing$, then $A^{\circ }$ $\cap$ $\partial B$ $\ne$ $\varnothing$ . Assume that $A^{\circ }$ $\cap$ $\partial B$ = $\varnothing$. Since B=$B^{\circ }$ $\cap$ $\partial B$. it follows that $\partial A$ $\cap$ B= $\varnothing$ and, therefore,B $\subset$ X - $\partial A$.Proposition 3.4 implies that $A^{\circ }$ and X -A form a separation of X - $\partial A$, and since B is connected, proposition 3.3 implies that either B $\subset$ $\partial A$ or B $\subset$ X - A . Since, by assumption, $\partial A$ - $\partial B$ $\ne$ $\varnothing$, it follows that B $\subset$ $A^{\circ }$ and, therefore, $\partial B$ $\subset$ $A^{\circ }$. Clearly, $\partial B$ $\cap$ $\partial A$ $\ne$ $\varnothing$ and the result follows.

证明了如果边界-边界交点为空而内-内交点为非空，则边界-内或内-边界交点均为非空。 这意味着具有非空的内部-内部交点和空的边界-边界、边界-内部和内部-边界的空间关系$r_2$不可能发生。 这样，命题的证明就完成了。 设A和B是$\partial A$ $\cap$ $\partial B$ $\ne$ $\varnothing$且$A^{\circ }$ $\cap$ $B^{\circ }$ $\ne$ $\varnothing$的空间区域。 当 $\partial A$ $\cap$ $B^{\circ }$= $\varnothing$时，则$A^{\circ }$ $\cap$ $\partial B$ $\ne$ $\varnothing$。 假设$A^{\circ }$ $\cap$ $\partial B$ = $\varnothing$。 因为B=$B^{\circ }$ $\cap$ $\partial B$。 因此$\partial A$ $\cap$ B= $\varnothing$，因此B $\subset$ X - $\partial A$。命题3.4暗示$A^{\circ }$和X -A构成X - $\partial A$的分离，而由于B是连接的，命题3.3暗示B $\subset$ $\partial A$或B $\subset$ X - A。 因为，根据假设，$\partial A$ - $\partial B$ $\ne$ $\varnothing$，所以 B $\subset$ $A^{\circ }$，因此$\partial B$ $\subset$ $A^{\circ }$。很明显，$\partial B$ $\cap$ $\partial A$ $\ne$ $\varnothing$，结果随之而来。

5.2 Semantics of reqion relations 区域关系的语义

In figure 1, examples weredepicted for the topological spatial relations 1 and 2 between spatial regions. Each of these nine relations is considered in the definitions belowand their semantics are investigated using the same notation as in Egenhofer (1989) and Egenhofer and Herring (1990).

在图1中，描述了空间区域之间的拓扑空间关系1和2的示例。 在下面的定义中考虑这九种关系中的每一种，并使用与Egenhofer(1989)和Egenhofer和Herring(1990)相同的符号来研究它们的语义。

Definition 5.4. The descriptive terms for the nine topological spatial relations between two regions are given in table 2.
定义5.4。 表2给出了两个区域之间的9种拓扑空间关系的描述术语。

If the topological spatial relation between A and B is '0 then, in the set-theoretic sense, A and B are disjoint and, therefore, the topological spatial relation disjoint coincides with the set-theoretic notion of disjoint. The following proposition and corollaries justify the other descriptive terms for the topological spatial relations defined in table 2.
如果A和B之间的拓扑空间关系为’0，则在集合意义上，A和B是不相交的，因此，不相交的拓扑空间关系与不相交的集合概念是一致的。 下面的命题和推论证明了表2中定义的拓扑空间关系的其他描述术语。

Table 2. Terminology used for the nine relations between two spatial regions.
表2. 指两个空间区域之间的九种关系的术语。

Proposition 5.5. Let A and B be spatial regions in X. If $A^{\circ }$ $\cap$ $B^{\circ }$ $\ne$ $\varnothing$ and $A^{\circ }$ $\cap$ $\partial B$ = $\varnothing$, then $A^{\circ }$ $\subset$ $B^{\circ }$ and A $\subset$ B.

命题5.5. 设A和B是x中的空间区域，如果$A^{\circ }$ $\cap$ $B^{\circ }$ $\ne$ $\varnothing$, $A^{\circ }$ $\cap$ $\partial B$ = $\varnothing$，则 $A^{\circ }$ $\subset$ $B^{\circ }$, A $\subset$ B.。

Proof: $A^{\circ }$ is connected. Proposition 3.4 implies that $B^{\circ }$ and X -B form a separation of X. Since $A^{\circ }$ $\cap$ $\partial B$ = $\varnothing$ , it follows by proposition 3.1 that $A^{\circ }$ < $B^{\circ }$ $\cup$ (X - B). Proposition 3.3 implies that either $A^{\circ }$ $\subset$ $B^{\circ }$ or $A^{\circ }$ $\subset$ (X-B). But $A^{\circ }$ $\cap$ $B^{\circ }$ $\ne$ $\varnothing$; therefore, $A^{\circ }$ $\subset$ {B^\circ}$. Since $A^{\circ }$ $\subset$ $B^{\circ }$, it follows that $\overline{A^{\circ }}$ $\subset$ $\overline{B^{\circ }}$ which, by definition 5.1, implies that A $\subset$ B .

证明: $A^{\circ }$ 连接。 命题3.4暗示了$B^{\circ }$和X -B构成了X的分离。 由于 $A^{\circ }$ $\cap$ $\partial B$ = $\varnothing$，因此命题3.1提出$A^{\circ }$ < $B^{\circ }$ $\cup$ (X - B)。命题3.3暗示要么 $A^{\circ }$ $\subset$ $B^{\circ }$，要么$A^{\circ }$ $\subset$ (X-B)。 但$A^{\circ }$ $\cap$ $B^{\circ }$ $\ne$ $\varnothing$; 因此，$A^{\circ }$ $\subset$ {B^{\circ}$，因为$\bar{A}\circ}$ $\subset$ $\overline{B^{\circ }}$，所以A < B，根据定义5.1，意味着 A $\subset$ B。

From proposition 5.5 it follows that if A.is covered by B, then A $\subset$ B; therefore, the spatial relation is covered by coincides with the set-theoretic notion of being a subset of.

根据命题5.5，如果A被B覆盖，那么A $\subset$ B; 因此，空间关系被集论中作为子集的集合概念所覆盖。

The following corollary to proposition 5.5 shows that the spatial relation equal corresponds to the set-theoretic notion of equality.

由下面命题5.5的推论可知，空间关系等于对应于集合论的相等概念。

Corollary 5.6. Let A and B be spatial regions. If the spatial relation between A and B is $r_3$, then A=B.

推论5.6. 设A和B为空间区域。若A与B的空间关系为 $r_3$ ，则A=B。

Proof: {A^\circ}$ $\cap$ {B^\circ}$ $\ne$ $\varnothing$ and {A^\circ}$ $\cap$ $\partial B$ = $\varnothing$;therefore, proposition 5.5 implies that A $\subset$ B Furthermore,$\partial A$ $\cap$ &{B^\circ}$ = $\varnothing$. Again by proposition 5.5, B $\subset$ A. Thus A=B.

证明: {A^\circ}$ $\cap$ {B^\circ}$ $\ne$ $\varnothing$,{A^\circ}$ $\cap$ $\partial B$ = $\varnothing$，因此，命题5.5暗示A $\subset$ B，进而，$\partial A$ $\cap$ {B^\circ}$ = $\varnothing$。 同样，根据命题5.5,B $\subset$ A，因此A=B。

The following corollary to proposition 5.5 shows that if A is inside B, then A $\subset$ $B^{\circ }$; therefore, the spatial relation inside coincides with the topological notion of being contained in the interior. Conversely, contains corresponds to contains in the interior.

下面命题5.5的推论表明，如果A在B内部，那么A $\subset$ $B^{\circ }$ ; 因此，内部的空间关系与包含在内部的拓扑概念相一致。 相反，contains对应于内部的contains。

Corollary 5.7. Let A and B be spatial regions. If the spatial relation between A and B is $r_6$ then A $\subset$ $B^{\circ }$.

推论5.7。 设A和B为空间区域。 如果A与B的空间关系为$r_6$，则A $\subset$ $B^{\circ }$。

Proof: Proposition 5.5 implies that $A^{\circ }$ $\subset$ $B^{\circ }$ and A $\subset$ B. By proposition 3.2, A= $A^{\circ }$ $\cup$ $\partial A$ and B= $B^{\circ }$ $\cup$ $\partial B$. So $\partial A$ $\subset$ $B^{\circ }$. Since $\partial A$ $\cap$ $\partial B$ = $\varnothing$, it follows that $\partial A$ $\subset$ $B^{\circ }$. Together with $A^{\circ }$ A$ $\subset$ $B^{\circ }$ this implies that A $\subset$ $B^{\circ }$.

证明:命题5.5 暗示 $A^{\circ }$ $\subset$ $B^{\circ }$, A $\subset$ B。命题3.2,A= $A^{\circ }$ $\cup$ $\partial A$, B= $B^{\circ }$ $\cup$ $\partial B$，所以 $\partial A$ $\subset$ $B^{\circ }$。因为 $\partial A$ $\cap$ $\partial B$= $\varnothing$，所以 $\partial A$ $\subset$ $B^{\circ }$，加上 $A^{\circ }$ A$ $\subset$ $B^{\circ }$，这意味着A $\subset$ $B^{\circ }$。

6. Relations in n-dimensional spaces n维空间中的关系

It is natural to ask ‘What further restrictions on the topological space X and the sets under consideration in X further reduce the topological spatial relations that can occur ?’ This section will explore this question by considering the case where X is a Euclidean space.

很自然地会问“对拓扑空间X和X中考虑的集合有什么进一步的限制进一步减少了可能发生的拓扑空间关系？”本节将通过考虑X是欧几里德空间的情况来探讨这个问题。

$R^n$ denotes n-dimensional Euclidean space with the usual Euclidean metric. A subset of $R^n$ is bounded if there is an upper bound to the distances between pairs of points in the set; otherwise, it is said to be unbounded.

$R^n$表示n维欧几里得空间，用通常的欧几里得度规。 如果集合中点对之间的距离有上界，则$R^n$的子集是有界的; 否则，它被称为无界。

The unit disk in $R^n$ is the set of points in $R^n$ whose distance from the origin is less than, or equal to, 1.The unit sphere in $R^n$ is the set of points in $R^n$ whose distance from the origin is equal to 1. For n $\ge$ 1 the unit disk in $R^n$ is connected. For n $\ge$ 2 the unit sphere in $R^n$ is connected. Let X be a topological space. An n-disk in X is a subspace of X that is homeomorphic to the unit disk in $R^n$. An n-sphere in X is a subspace of X that is homeomorphic to the unit sphere in $R^{n+1}$. n-disks in $R^n$ are bounded and are spatial regions; the latter is a relatively straightforward consequence of the Brouwer theorem on the invariance of domain (Spanier 1966). Since n-disks in $R^n$ are spatial regions, proposition 5.3 restricts the number of spatial relations that can occur between them.

$R^n$中的单位圆盘是$R^n$中到原点的距离小于等于1的点的集合。 $R^n$中的单位球面是$R^n$中的点的集合它们到原点的距离是1。 对于n $\ge$ 1，连接$R^n$的单元硬盘。 对于n $\ge$ 2, $R^n$中的单位球是连通的。 设X是一个拓扑空间。 X中的n盘是X的一个子空间，它是$R^n$中的单位盘的同胚。 X中的n球是X的一个子空间，它是$R^{n+1}$中的单位球的同胚。 $R^n$中的n个磁盘是有界的，是空间区域; 后者是Brouwer定理关于定义域不变性的一个相对简单的推论(Spanier 1966)。 由于$R^n$中的n盘是空间区域，命题5.3限制了它们之间可能发生的空间关系的数量。

In proposition 6.1 it is shown that if A and B are n-disks in $R^n$ with n$\geq$2, then the spatial relation overlap with disjoint boundary cannot occur. The proof of this proposition is based on the following two facts:

在命题6.1中，如果A和B是 $R^n$中的n个盘，且n$\geq$2，则不可能发生边界不相交的空间关系重叠。 该命题的证明基于以下两个事实:

Fact 1. Let A be an n-disk in $R^n$ with n $\ge$ 2. Then $\partial A$ is an (n-1)-sphere in $R^n$ and, therefore, connected.
事实1. 设A是 $R^n$ 中的n盘n $\ge$ 2。 那么 $\partial A$ 是 $R^n$ 中的(n-1)球面，因此是连通的。

This fact, also, is a consequence of the Brouwer theorem on the invariance of domain (Spanier 1966).
这个事实也是关于定义域不变性的Brouwer定理的一个结果(Spanier 1966)。

Fact 2. Let A be an n-disk in $R^n$ with n $\ge$ 2. Then $R^n$ - $A^{\circ }$ is connected and unbounded.
事实2。 设A是 $R^n$ 中的n盘n $\ge$ 2。 那么 $R^n$ - $A^{\circ }$ 是连通的无界的。

This second fact is a (non-)separation theorem related to the Jordan-Brouwer separation theorem (Spanier 1966).
第二个事实是一个与Jordan-Brouwer分离定理(Spanier 1966)相关的(非)分离定理。

Proposition 6.1. The topological spatial relation $r_{14}$ , overlap with disjoint boundaries, does not occur between n-disks in R with n $\ge$ 2.
命题6.1。 在n $\ge$ 2中，n个盘之间不存在与不相交边界重叠的拓扑空间关系 $r_{14}$。

Proof: Let A and B be n-disks in $R^n$ with n $\ge$ 2. It is shown that if $\partial A$ $\cap$ $\partial B$ = $\varnothing$, then A and B do not overlap and, therefore, the spatial relation overlap with disjoint boundaries cannot occur.

证明:设A和B是 $R^n$ 中的n个，n $\ge$ 2。 结果表明，如果 $\partial A$ $\cap$ $\partial B$ = $\varnothing$，则A和B不重叠，因此不存在边界不相交的空间关系重叠。

Assume $\partial A$ $\cap$ $\partial B$ = $\varnothing$ and A and B overlap. A contradiction will be derived. B is a spatial region; therefore, proposition 3.4 implies that $B^{\circ }$ and $R^n$ - B form a separation of $R^n$ - $\partial B$. As $\partial A$ - $\partial B$= $\varnothing$ it follows that $\partial A$ $\subset$ $R^n$ - $\partial B$.

假设 $\partial A$ $\cap$ $\partial B$ = $\varnothing$, A和B重叠。 这样就会产生矛盾。 B是一个空间区域; 因此，命题3.4暗示 $B^{\circ }$ 和 $R^n$ - B构成了 $R^n$ - $\partial B$ 的分离。当 $\partial A$ - $\partial B$= $\varnothing$ 时，可以得出 $\partial A$ $\subset$ $R^n$ - $\partial B$。

By fact 1, $\partial A$ is connected, therefore, proposition 3.3 implies that either $\partial A$ $\subset$ $B^{\circ }$ or $\partial A$ $\subset$ ($R^n$ - B). Since A and B overlap, it follows that A-B=0 and, therefore, $\partial A$ $\subset$ $B^{\circ }$ .

根据事实1, $\partial A$ 是连通的，因此，命题3.3意味着 $\partial A$ $\subset$ $B^{\circ }$ 或 $\partial A$ $\subset$ ($R^n$ -B).由于A和B重叠，因此 $\partial A$ $\cap$ $B^{\circ }$ $\ne$ $\varnothing$，因此 $\partial A$ $\subset$ $B^{\circ }$。

$\partial A$ $\subset$ $B^{\circ }$ implies that $\partial A$ $\cap$ ($R^n$ - $B^{\circ }$)= $\varnothing$ . By fact 2, $R^n$ - $B^{\circ }$ is connected. Using propositions 3.3 and 3.4 and arguing as above, it follows that either ( $R^n$ - $B^{\circ }$ ) $\subset$ $A^{\circ }$ or ($R^n$ - $B^{\circ }$ )$\subset$($R^n$ -A). The first case yields a contradiction because, by fact 2, $R^n$ - $B^{\circ }$ is unbounded, but $A^{\circ }$ is not.

$\partial A$ $\subset$ $B^{\circ }$ 意味着 $\partial A$ $\cap$ ($R^n$ - $B^{\circ }$)= $\varnothing$ 。 根据事实2, $R^n$ - $B^{\circ }$ 是连通的。 使用命题3.3和3.4并如上论证，可以得出( $R^n$ - $B^{\circ }$ ) $\subset$ $A^{\circ }$ 或($R^n$ - $B^{\circ }$ ) $\subset$ ($R^n$ -A)。 第一种情况产生了矛盾，因为，事实上，$R^n$ - $B^{\circ }$ 是无界的，而 $A^{\circ }$ 不是。

The second case implies that A $\subset$ $B^{\circ }$ and, therefore, $\partial A$ $\cap$ $B^{\circ }$ = $\varnothing$, which contradicts the assumption that A and B overlap. Therefore, in either case a contradiction is obtained and it followsthat the spatial relation $r_{14}$ cannot occur between n-disks in $R^n$ with n $\ge$ 2.

第二种情况意味着 A $\subset$ $B^{\circ }$，因此，$\partial A$ $\cap$ $B^{\circ }$ = $\varnothing$，这与A和B重叠的假设相矛盾。 因此，在任何一种情况下，都得到了一个矛盾，并由此得出，在n $\ge$ 2中，在$R^n$中的n个盘之间不存在空间关系$r_{14}$。

Note that for n $\ge$ 2 the topological spatial relation r overlap with intersecting boundaries, does occur between two n-disks (figure 1).

注意，对于n $\ge$ 2，拓扑空间关系r与相交边界重叠，确实发生在两个n个磁盘之间(图1)。

The opposite situation occurs in $R^1$ where r can occur between 1-disks, while $r_{15}$, overlap with intersecting boundaries, cannot. It is clear that $r_{14}$ can occur between two 1-disks in $R^1$(figure 2). Proposition 6.2 shows that $r_{15}$ cannot occur. Its proof requires the easily derived fact that a spatial region in $R^1$ is either a closed interval [a, b] for some a,b $\in$ $R^1$, or a closed ray [a, $\infty$) or (-$\infty$,a] for some a $\in$ $R^1$.

相反的情况发生在$R^1$, r可以出现在1-disks中，而$r_{15}$，与交叉边界的重叠，不能出现。 显然，$r_{14}$可以出现在$R^1$中的两个1盘之间(图2)。命题6.2表明$r_{15}$不能出现。 它的证明需要很容易推导出这样一个事实:对于某些a,b $\in$ $R^1$, R中的空间区域是一个闭区间[a, b]，或者对于某些a $\in$ $R^1$，是一个闭射线[a, $\infty$)或(-$\infty$,a]。

Proposition 6.2. The topological spatial relation $r_{15}$ does not occur between spatial regions in $R^1$.
命题6.2. 在 $R^1$ 中，空间区域之间没有拓扑空间关系 $r_{15}$。

Proof: Let A and B be spatial regions in $R^1$ and assume that A and B overlap. It is shown that $\partial A$ $\cap$ $\partial B$ =$\varnothing$. Each of A and B is a closed interval or a closed ray; therefore, there are nine different cases to examine. One is selected; the others can be proven "accordingly.
证明:设A、B为 $R^1$ 中的空间区域，假设A与B重叠。 得到 $\partial A$ $\cap$ $\partial B$ =$\varnothing$。 A和B的每一个都是一个闭区间或闭射线; 因此，有九个不同的案例需要研究。 一个被选中; 其他的可以“相应地”证明。

Assume A =[a, $\infty$) and B=(-$\infty$,a]. Then A ={a} and B= {b}. Since A and B overlap, it follows that a< b, which implies that $\partial A$ $\cap$ $\partial B$ =$\varnothing$.

假设A =[a, $\infty$)， B=(-$\infty$,a]。 那么A ={a}， B= {b}。 因为A和B重叠，所以A < B，这意味着 $\partial A$ $\cap$ $\partial B$ =$\varnothing$。

7. Conclusion 结论

A framework for the definition of topological spatial relations has been presented.It is based on purely topological properties and is thus independent of the existence of a distance function. The topological relations are described by the four intersections of the boundaries and interiors of two point-sets. Considering the binary values empty and non-empty for these intersections, a set of 16 mutually exclusive specifications has been identified. Fewer relations exist if particular restrictions on the point-sets and the topological space are made. It was proved that there are only nine topological spatial relations between point-sets which are homeomorphic to polygonal areas in the plane.

提出了一种拓扑空间关系的定义框架。 它基于纯粹的拓扑性质，因此与距离函数的存在无关。 拓扑关系用两个点集的边界和内部的四个交点来描述。 考虑到这些交集的二进制值为空和非空，确定了一组16个互斥规范。 如果对点集和拓扑空间作特殊的限制，则其关系较少。 证明了平面上与多边形区域同胚的点集之间只有9个拓扑空间关系。

Although the nature of this work is rather theoretical, the framework has an immediate effect on the design and implementation of geographic information systems. Previously, for every topological spatial relation a separate procedure had to be programmed and no mechanism existed to assure completeness. Now, topological spatial relations can be derived from a single, consistent model and no programming for individual relations will be necessary. Prototype implementations of this framework have been designed and partially implemented (Egenhofer 1989), and various extensions to the framework have been investigated to provide more details about topological spatial relations, such as the consideration of the dimensions of the intersections and ofthe number of disconnected subsetsin the intersections (Egenhofer and Herring 1990).Ongoing investigations focus on the application of this framework for formal reasoning about combinations of topological spatial relationships.

虽然这项工作的性质是相当理论性的，但该框架对地理信息系统的设计和实施有直接的影响。 以前，对于每一个拓扑空间关系，都必须编写一个单独的程序，并且没有机制来保证完整性。 现在，拓扑空间关系可以从一个单一的、一致的模型中得到，并且不需要对单个关系进行编程。 该框架的原型实现已经设计并部分实现(Egenhofer 1989)，并且对框架的各种扩展进行了研究，以提供关于拓扑空间关系的更多细节， 例如考虑交叉口的维数和交叉口中未连接子集的数量(Egenhofer和Herring 1990)。 正在进行的研究集中在该框架的应用，以进行拓扑空间关系组合的形式推理。

The framework presented is considered a start and further investigations are necessary to verify its suitability. Here, only topological spatial relations with codimension zero were considered, i.e. the difference between the dimension of the space and the dimension of the embedded spatial objects is zero, e.g. between regions in the plane and intervals on the one-dimensional line, Also of interest for GIS applications are the topological spatial relationships with co-dimension greater than zero, e.g. between two lines in the plane (Herring 1991). Likewise, the applicability of this framework to topological spatial relations between objects of different dimensions, such as a region and a line, must be tested.

提出的框架被认为是一个开始，进一步的调查是必要的，以验证其适用性。 这里只考虑余维为零的拓扑空间关系，即空间维数与嵌入空间对象维数之差为零，如平面内区域与一维直线上区间之间， GIS应用的另一个兴趣点是余维大于零的拓扑空间关系，例如平面上两条线之间的关系(Herring 1991)。 同样，该框架对于不同维度对象(如区域和直线)之间的拓扑空间关系的适用性也必须进行测试。

Acknowledgments 致谢

The motivation for this work was given by Bruce Palmer. During many discussions with John Herring, these concepts have been clarified. Andrew Frank and Renato Barrera made valuable comments on an earlier version of this paper. This work was partially funded by grants from NSF under No. IST 86-09123, Digital Equipment Corporation under Sponsored Research Agreement No. 414 and TP-765536, Intergraph Corporation, and the Bureau of the Census under Joint Statistical Agreement N o. 89-23. Additional support from NSF for the NCGIA under grant number SES 88-10917 is gratefully acknowledged.

这项工作的动机由Bruce Palmer给出。 在与约翰·赫林的多次讨论中，这些概念已经得到了澄清。 安德鲁·弗兰克和雷纳托·巴雷拉对本文的早期版本作了有价值的评论。 这项工作的部分资金来自美国国家科学基金会No。 IST 86-09123，数字设备公司，根据赞助研究协议No. 414和TP-765536，鹰图公司，和人口普查局根据联合统计协议No. 89-23。 感谢NSF对NCGIA的额外支持，资助编号为SES 88-10917。

References 引用

ABLER, R., 1987, TheNational Science Foundation National CenterforGeographic Infonnation and Analysis. International Journal of Geographical Information Systems, I, 303-326.

ALLEN, J. F., 1983, Maintaining knowledge about temporal intervals. Communications of the ACM, 26, 832-843.

CHANG, S.K., JUNGERT, E.,and LI, Y 1989, The design of pictorialdatabases based upon the theory of symbolic projections. In: Proceedings of Symposium on the Design and Implementation of Large Spatial Databases held in Santa Barbara, CA, edited by A. Buchmann, O. Giinthet, T. Smith, and Y. Wang (New York: Springer-Verlag), Lecture Notes in Computer Science, Vol. 409, pp.303-323.

CLAIRE, R~ and GUPTILL, S~ 1982, Spatialoperators forselected data structures.In: Proceedings ofAuto-Carto V. held in Crystal City, Virginia, pp. 189-200.

EGENHOFER, M., 1989, A formal definition of binary topological relationships. In: Third lmemational Confereru:e on FourrdDtions of Data Organization and Algorithms (FODO), held in 'Paris, France;edited by W. Litwin and H.-J. Schek (New York: Springer-Verlag), Lecture Notes in Computer Science, Vol. 367, pp.451-472.

EGENHoFER, M~ and HERRING, J., 1990, A mathematical framework for the definition of topological relationships. In: Proceedings ofFourth InternationDI Symposium on Spotial Data Handling held in Zurich, Switzerland, edited by K. Brassel and H. Kishimoto,pp. 803-813.

FRANK, 1982, Map query-e-database query language.for retrieval of geometric data and its graphical representation. ACM Computer Graphics, 16, 199-201.

FREEMAN, J., 1915, The modelling ofspatial relations. Computer Graphics and Image Processing,4, 156-111.

GOTiNG. R., 1988, Gee-relational algebra: a model and query language for geometric database systems. In: International Conference on Extending Database Technology held in VeniCe, Italy, edited by J. Schmidt, S. Ceri, and M. Missikolf(New York: Springer-Verlag), Lecture Notes in Computer Science, Vol. 303, pp.506-521.

HERNANDEZ, D., 1991, Relative representation ofspatial knowledge: the 2-D case. In: Cognitive and Linguistic Aspects ofGeographic Space, edited by D. Mark and A. Frank (Dordrecht:Kluwer Academic) (in press),

HERRING, J~ 1991, The mathematical modeling of spatial and non-spatial information in geographic information systems. In: Cognitive and Linguistic Aspects ofGeographic Space,edited by D. Mark and A. Frank (Dordrecht: Kluwer Academic) (in press).

HERRING, J., LARSEN, R~ and SHiVAKUMAR, J~ 1988, Extensions to the SQL language to support spatial analysis in a topological database. In: Proceedings of GIS/LIS 'SS, held in San Antonio, Texas, pp.741-15O.

HERSKOVlTS, A., 1986, Language and Spatwl Cognition–An Interdisciplinary Study of the Prepositions in English (Cambridge: Cambridge University Press).

INGRAM, K~ and PHILUI’S, W., 1981,'Geographic information processing using a SQL.based query language. In: Proceedings ofAUTO-CARTO S, Eighth InternationDISymposium on Computer-Assisted Cartography held in Baltimore, Maryland, edited by N. R. Chrisman,pp.326-335.

KAINZ, W., 1990, Spatial relationships-topology versus order. In: Proceedings of Fourth InternationDISymposium on Spatial Data Handling held in Zurich, Switzerland, edited by K.Brassel and H. Kishimoto, pp.814-819.

MOlENAAR, M., 1989,Single valued vector maps-a concept in geographic information systems.Geo-lrformauonssysteme.L; 18-26.

MUNKRES, J., 1966, Elemenatary Differential Topology (Princeton, NJ: Princeton University Press).

NATIONAL CENTER FOR GEOGRAPHIC INFORMATION AND ANALYSIS (NCGIA), 1989,The research plan of the National Center for Geographic Information and Analysis. International Journal of Geographical Information Systems, 3, 117-136.

PEUQUET, D., 1986, The use of spatial. relationships to aid spatial database retrieval. In:Proceedings ofSecond International Symposium on Spatial Data Handling held in Seattle, WA, edited by D. Marble, pp.459-411.

PEUQUET, D., and CI-XIANG, Z, 1981, An algorithm to determine the directional relationship between arbitrarily-shaped polygons in the plane. Pauern Recognition, 20, 65-14.

PULLAR, D., 1988, Data definition and operators on a spatial data model. In: Proceedings of ACSM-ASPRS Annual Convention, held in St. Louis, Missouri, pp. 191-202.

PULLAR, D., and EGENHOFER, M., 1988, Towards formal definitions of topological relations among spatial objects. In: Proceedings of Third International Symposium on Spatwl Data Handling held in Sydney, Australi4, edited by D. Marble, pp.225-242.

Roussosoutos, N., FALOUTSOS, C~ and SaLIS, T., 1988, An efficient pictorial database system for PSQL. IEEE Transactions on Software Engineering, 14, 63()-{)38.

SMITH, T., PEUQUET, D~ MENON, S., and AGRAWAL, P., 1981, KBGIS-II: a knowledge-based geographical information system. International Journal of Geographical Information Systems, 1, 149-112.

SPANIER, E., 1966, Algebraic Topology (New York: McGraw-Hili).

TALMY, L., 1983, How language structures space. In: Spatial Orientation: Theory, Research, and Application, edited by H. Pick and L. Acredolo (New York: Plenum Press), pp. 225-282.

An efficient pictorial database system for PSQL. IEEE Transactions on Software Engineering, 14, 63()-{)38.

SMITH, T., PEUQUET, D~ MENON, S., and AGRAWAL, P., 1981, KBGIS-II: a knowledge-based geographical information system. International Journal of Geographical Information Systems, 1, 149-112.

SPANIER, E., 1966, Algebraic Topology (New York: McGraw-Hili).

TALMY, L., 1983, How language structures space. In: Spatial Orientation: Theory, Research, and Application, edited by H. Pick and L. Acredolo (New York: Plenum Press), pp. 225-282.

WAGNER, D~ 1988, A method of evaluating polygon overlay algorithms. In: Proceedings of ACSM-ASPRS Annual Convention, held’in St. Louis, Missouri. pp.113-183.