Journal of Official Statistics, Vol.4, No.2, 1988. pp. 125–140
Classifying and Comparing Spatial Relations of Computerized Maps for Feature Matching Applications
Abstract:Modern computerized maps either contain digital information on spatial relations, such as adjacency relations, shape, network patterns, and measures of position and distance of features, or they permit derivation of that information from the feature data that they do contain. Such spatial attributes lend themselves to computerized statistical analysis much like any other data. Comparative data analysis of spatial relations is possible when two map files are known to cover the same area. In this case, spatial characteristics alone may be used to establish linkages between many of the feature records of the two files. This paper presents examples of some spatial measures of distance and local configuration that were used to develop an automated feature matching system at the U.S. Bureau of the Census. For a particular sample pair of maps, global summaries and spatial depictions of distance and configuration measures are presented; and some additional uses for the measures are suggested.
Keywords:Computerized maps; map distortion; automated cartography; feature matching; record linkage; configuration; conflation; spider function.
Copyright © Statistics Sweden 1996-2017. Open AccessISSN 0282-423XCreated and Maintained by OKS Group