Using stroke and mesh to recognize building group patterns

ABSTRACT

Building patterns are crucial structures and should be preserved in map generalization. However, while building patterns are not explicitly described in building datasets, map readers perceive building patterns effortlessly. Hence, to better support map generalization, it is important to automatically recognize building patterns in such datasets. This paper first proposes an extended and integrated typology of different building patterns. Based on the typology, building patterns are recognized using stroke and mesh. This method first structures the proximity graph of buildings, and then introduces six constraints (distance, size, shape, orientation, elongation, and facing ratio) to refine the original proximity graph. Strokes and meshes are derived from the refined proximity graph, and are used to recognize linear and grid building patterns, respectively. The proposed method is tested in four regions that are representative of different pattern types. The recognition results are evaluated in an expert survey and compared with the minimum spanning tree method. Assessment suggests that the linear and grid patterns in suburban and rural areas are recognized with satisfying results.

RÉSUMÉ

Les modèles de construction sont des structures cruciales et doivent être préservés dans la généralisation de la carte. Cependant, bien que les modèles de construction ne soient pas explicitement décrits dans les jeux de données de construction, les lecteurs de carte perçoivent les modèles de construction sans effort. Par conséquent, pour mieux prendre en charge la généralisation de la carte, il est important de reconnaître automatiquement les modèles de construction dans ces jeux de données. Ce document propose d’abord une typologie étendue et intégrée de différents modèles de construction. Sur la base de la typologie, les modèles de construction sont reconnus à l'aide de traits et de maillages. Cette méthode structure d'abord le graphe de proximité des bâtiments, puis introduit six contraintes (distance, taille, forme, orientation, allongement et rapport de parement) pour affiner le graphe de proximité d'origine. Les traits et les maillages sont dérivés du graphe de proximité affiné et sont utilisés pour reconnaître les modèles de construction linéaire et en grille, respectivement. La méthode proposée est testée dans quatre régions représentatives de différents types de modèles. Les résultats de reconnaissance sont évalués dans le cadre d'une enquête auprès d'experts et comparés à la méthode du spanning tree minimum. L’évaluation suggère que les schémas linéaires et quadrillés dans les zones suburbaines et rurales sont reconnus avec des résultats satisfaisants.

KEYWORDS:

• Generalization
• building grouping
• building pattern recognition
• stroke
• network mesh

Acknowledgement

Comments from the editor and anonymous reviewers are greatly appreciated.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes on contributors

Xiao Wang is a PhD student in the Institute of Cartography at Dresden University of Technology. He got his Master Degree in Engineering in China 2012. His master thesis is about spatial data matching. Currently, his research interests in automated map generalization, especially on the building group pattern detection and building generalization.

Dirk Burghardt is a full professor in the Cartographic Institute at Dresden University of Technology since 2009. From 1998 to 2001 he worked as a software engineer and product manager in a map production company in Switzerland. Between 2002 and 2008 he was lecturer in the GIS Division at the University of Zurich (CH). Professor Burghardt was the chair of the ICA Commission on Generalisation and Multiple Representation of the International Cartographic Association from 2011 to 2017. His main research interests include automated generalization and map production, geographic information retrieval, interactive cartographic presentations, geovisual analytics and cartographic communication.

ORCID

Xiao Wang https://orcid.org/0000-0002-4666-0392

Dirk Burghardt http://orcid.org/0000-0003-2949-4887

Additional information

Funding

This work was supported by Deutsche Forschungsgemeinschaft [Grant Number BU 2605/5-1, WPS-PolyGen].