Template Matching in Support of Generalisation of Rural Buildings

A range of methodologies has been proposed to derive generalised forms of buildings at coarser scales. In this paper, we explore the use of simple pattern matching algorithms in order to select from a set of templates, a building outline that best characterises a more detailed form. This template matching process is used to simplify the form of rural buildings (farmsteads) in Danish mapping, and is currently a manual task within the National Mapping Agency of Denmark (KMS). This research has explored the feasibility of automating this approach, and reports on its implementation, and provides an evaluation. The challenge in pattern matching is to minimise the misidentification of patterns (type I error) and failures to find any match (type II error) in the assignment of templates. The initial set of results was very encouraging.

[1]  Remco C. Veltkamp,et al.  Shape matching: similarity measures and algorithms , 2001, Proceedings International Conference on Shape Modeling and Applications.

[2]  Remco C. Veltkamp,et al.  State of the Art in Shape Matching , 2001, Principles of Visual Information Retrieval.

[3]  Robert Weibel,et al.  Generalising spatial data and dealing with multiple representations , 1999 .

[4]  Dominique Attali,et al.  Computing and Simplifying 2D and 3D Continuous Skeletons , 1997, Comput. Vis. Image Underst..

[5]  William Mackaness,et al.  Dynamic generalisation from single detailed database to support web based interaction , 1999 .

[6]  Terry Caelli,et al.  Building Detection Using Bayesian Networks , 2000, Int. J. Pattern Recognit. Artif. Intell..

[7]  Sven Loncaric,et al.  A survey of shape analysis techniques , 1998, Pattern Recognit..