Spatial Relations between Classes as Integrity Constraints

Tarquini, Francesco; Clementini, Eliseo

doi:10.1111/j.1467-9671.2008.01134.x

Cited by 9 publications

(4 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Spatial operators that are invoked by functions can be distinguished in those already implemented in OGC standards, such as topological relations, Euclidean distance, set operations, convex hull, and many others, and those for which it does not exist a standard implementation: regarding this latter group, we implemented our own version of operators (some of them are listed in Table 1). For more information on spatial operators, it is possible to refer to the broad literature on them, such as topological relations [20,21], projective relations [22]- [26], directional relations [27], and visibility relations [28]- [31].…”

Section: B Methodologymentioning

confidence: 99%

Extraction of Land Cover Units from Land Cover Components Based on Geometric Rules

Clementini

Natali

2019

Int. J. Adv. Sci. Eng. Inf. Technol.

Self Cite

View full text Add to dashboard Cite

Land cover units are aggregations of land cover components that are obtained by using criteria of homogeneity and proximity of basic components. For example, residential urban settlements can be defined as aggregations of single buildings, neighboring green spaces, paved surfaces and small roads, which are separated by more prominent land cover components, such as main roads or rivers. Land cover components belong to standard classes typically obtained by an automated classification process applied to aerial or satellite images, such as buildings, constructed areas, bare soil, water, vegetation, and the like. Land cover units belong to more general classes, obtained by a combination of land cover components, such as residential areas, industrial areas, road networks, river systems, and agricultural units. In this paper, we describe an approach based on the application of geometric rules and semantic constraints to extract land cover units from land cover components. We use spatial operators to extract composite land cover units from land cover databases, where spatial operators are taken from standards of the Open Geospatial Consortium. Expert knowledge needs to be translated into specific automatic procedures, called complex object definitions or CODs. Finally, we build a prototype system, where the user can choose among a set of available CODs to build a sequence of actions that lead to the discovery of knowledge. We discuss several study cases, such as the recognition of urban settlements, agricultural land units, and road networks.

show abstract

Section: B Methodologymentioning

confidence: 99%

Extraction of Land Cover Units from Land Cover Components Based on Geometric Rules

Clementini

Natali

2019

Int. J. Adv. Sci. Eng. Inf. Technol.

Self Cite

View full text Add to dashboard Cite

show abstract

“…The modeling of uncertainty in topological relations is not a new topic, though the way we approach it in this article is new. The study of spatial relations when the objects are “crisp” (i.e., their boundary can be fully described by a one‐dimensional line) is an established subfield of GIS (Bartie, Clementini, & Reitsma, 2013; Clementini, 2008, 2013, 2019; Clementini & Cohn, 2014; Clementini, Skiadopoulos, Billen, & Tarquini, 2010; Fogliaroni & Clementini, 2014; Tarquini & Clementini, 2008). A stage of research about uncertainty in spatial data has concentrated on the adoption of “broad” boundaries in place of crisp boundaries; this has led to the definition of new spatial objects and a category of approximate topological relations (e.g., “nearly touch”) (Clementini, 2005; Clementini & Di Felice, 2001).…”

Section: State Of the Artmentioning

confidence: 99%

Homological relations: A methodology for the certification of irregular tessellations

Clementini

Lejdel²,

Mazzagufo

et al. 2020

Transactions in GIS

Self Cite

View full text Add to dashboard Cite

In GIS, spatial analysis is based on the use of spatial operations such as testing the spatial relations between features. Often, such tests are invalidated by errors in datasets. It is a very common experience that two bordering regions which should obey the topological relation “meet” fall instead in the “overlap” category. The situation is exacerbated when applying topological operators to regions that come from different datasets, where resolution and error sources are different. Despite the problem being quite common, up to now no standard approach has been defined to deal with spatial relations affected by errors of various origins. Referring to topological relations, we define a model to extend the eight Egenhofer relations between two simple regions: we call them homological relations (H‐relations). We discuss how exact topological relations can be extracted from observed relations and discuss the case of irregular tessellations, where errors have the most impact on vector data. In the proposed case study within the domain of geographic crowdsourced data, we propose algorithms for identifying homological regions and obtaining a corrected tessellation. This methodology can be considered as a step for quality control and the certification of irregular tessellations.

show abstract

“…An object at the user level (for example, a stream) can have multiple representations at the geometric level, since a river can be represented as a single line or a complex line, or as a two-dimensional region. Therefore, a semantic spatial relations between two streams (for example, one river flows into another watercourse) can be modelled with various geometric relations based on the adopted representations [63]. There are also architectural problems on how to interpret ontology at the user level in terms of data type and operators at the computational level [17].…”

Section: Levels Of Representationmentioning

confidence: 99%

A Conceptual Framework for Modelling Spatial Relations

Clementini¹

2019

ITC

Self Cite

View full text Add to dashboard Cite

Various approaches lie behind the modelling of spatial relations, which is a heterogeneous and interdisciplinary field. In this paper, we introduce a conceptual framework to describe the characteristics of various models and how they relate each other. A first categorization is made among three representation levels: geometric, computational, and user. At the geometric level, spatial objects can be seen as point-sets and relations can be formally defined at the mathematical level. At the computational level, objects are represented as data types and relations are computed via spatial operators. At the user level, objects and relations belong to a context-dependent user ontology. Another way of providing a categorization is following the underlying geometric space that describes the relations: we distinguish among topologic, projective, and metric relations. Then, we consider the cardinality of spatial relations, which is defined as the number of objects that participate in the relation. Another issue is the granularity at which the relation is described, ranging from general descriptions to very detailed ones. We also consider the dimension of the various geometric objects and the embedding space as a fundamental way of categorizing relations.

show abstract

Spatial Relations between Classes as Integrity Constraints

Cited by 9 publications

References 7 publications

Extraction of Land Cover Units from Land Cover Components Based on Geometric Rules

Extraction of Land Cover Units from Land Cover Components Based on Geometric Rules

Homological relations: A methodology for the certification of irregular tessellations

A Conceptual Framework for Modelling Spatial Relations

Contact Info

Product

Resources

About