Comparing the vario-scale approach with a discrete multi-representation based approach for automated generalisation of topographic data

Datasets at different scales can be obtained and maintained by two main approaches that differ considerably. The first option is the vario-scale approach that offers the possibility to derive a map at an arbitrary map scale (hence its name) by performing once a fully automatic generalisation process on the input data set. This sounds as the optimal situation. However, since fast and fully automated generalisation of topographic data is still a challenge, National Mapping Agencies (NMAs) employ a second option, i.e. the multi-scale approach, to produce and maintain maps at different, predefined scales. Examples are reported in the literature for IGN, France (Lecordix et al. 2007), KMS, Denmark (Foerster et al, 2010), ICC, Catalonia (Baella & Pla 2005), Germany (AdV, 2007), and The Netherlands (Stoter et al, 2009). The tGAP data structure (van Oosterom, 1995; van Oosterom, 2005; Meijers, 2011) provides a promising implementation of the vario-scale approach in which the most detailed data is stored once, and an incremental object by object generalisation process is run and represented in a data structure, which can afterwards be used to efficiently obtain any arbitrary scale on the fly (Vermeij et al, 2003; Meijers, 2011). Figure 1 shows the concept of tGAP.