Qualitative Representations of Schematized and Distorted Street Segments in Sketch Maps

Abstract

Over the past years, user-generated content has gained more prominence in the area of geographic information science. In many geo-spatial applications, sketch maps are considered as an intuitive user interaction modality. As typically only qualitative relations are persevered in sketch maps, processing spatial information on qualitative level has been suggested. The information represented in sketch maps are distorted, schematized, incomplete, and generalized. Particularly, streets are schematized and their angles of connectivity at junctions are distorted. In sketch maps, streets are lineal features and each street consists of one or more connected street segments. Straightening curved streets and distorting angles affect the qualitative information of streets. Thus, processing spatial information from sketch maps and making it available in information systems requires reliable formalizations which are robust against cognitive distortions. In this paper, we proposed a set of plausible representations to formalize the qualitative information of street segments in street networks. Further, we identify suitable orientation sectors. The orientation sectors are scrutinized and encoded with the proposed representation to capture the relative orientations of street segments. Afterwards, the representations are evaluated for the qualitative alignment of street segments from sketch maps with geo-referenced maps. The successful alignment of street segments will allow us to integrate spatial information from sketch maps into a Geographic Information System (GIS) as volunteered informationKeywordsSketch map alignmentQualitative constraint networksStreet segmentsStreet networkOrientation sectorsQualitative representation