Developing a vario-scale IMGeo using the constrained tGAP structure

Developing a vario-scale IMGeo using the constrained tGAP structure

Hofman, A.M.

Van Oosterom, P.J.M. (mentor)
Dilo, A. (mentor)
van Loenen, B. (mentor)

OTB Research Institute for the Built Environment

GIS technology

2008-02-19

The aim of this master thesis has been to answer the question: How can a vario-scale IMGeo be designed and developed by applying the constrained tGAP structure with Top10NL as initial constraint? This research has been done under authority of Gemeentewerken Rotterdam and TU Delft. For Gemeentewerken Rotterdam the main goal is to investigate the possibilities of automatic generalisation based on their project "Basisregistratie Geografie", for TU Delft the main goal is to see the concept of the constrained tGAP structure being investigated. The models IMGeo, a new large scale topographical standard in The Netherlands, and Top10NL, the authentic registration for medium scale topography in The Netherlands, are both derived from NEN 3610; they form the basis for the generalisation in this research. The main differences are explained by their backgrounds. IMGeo originates from the GBKN, whereas Top10NL is based on Top10Vector. The fact that these models are not made cooperatively is bad luck for this research; no object class hierarchy could be made from these models, because there are object classes in Top10NL which don't even exist in IMGeo. The Top10NL object classes therefore can't be seen as a generalisation of the object classes of IMGeo. The generalisation approach in this research was the constrained tGAP structure, a concept from the universities of Hannover and Delft. In this research IMGeo objects were assigned to Top10NL regions. Four possible methods to assign IMGeo objects to Top10NL regions were developed: - Simple overlay method. An intersection between the models where every IMGeo object is split at the borders of the overlapping Top10NL object. In the end result only Top10NL geometry will be visible. - The maximum area method. The Top10NL object which overlaps the IMGeo object the most is the shape to which the whole IMGeo object is assigned to. The IMGeo geometry is kept in this method. - The 35%-split method. If an IMGeo object belongs for more than 35% to two Top10NL objects we consider this Top10NL geometry as enrichment of the structure; therefore the IMGeo object is split and a new IMGeo object is created. For all IMGeo objects that don't have two Top10NL objects overlapping for more than 35% the maximum area method is applied. - The building first method. This method assigns IMGeo-buildings to a building region in case of some overlap with a Top10NL building without considering the amount of overlap. The other IMGeo objects are selected as in the maximum area method. The building first method has been developed because the results of the maximum area method and the 35%-split method were unsatisfying. This building first method gives the best results of the four methods. With this method further research was done. For the test dataset from the municipality of Almere class weights and class compatibilities have been derived, which are input values for the tGAP structure. This has been done after tuning the weights and compatibilities of the constrained tGAP structure and comparing the end result to the Top10NL dataset. Final conclusion of this report is that the constrained tGAP absolutely offers possibilities for automatic generalisation from large to medium scale data. However, the amount of pre-processing of the data and the state of development of the tGAP structure are reasons for a governmental organisation like Gemeentewerken Rotterdam to not yet develop a product like this. A topological structure of the models might possibly solve these problems. Cooperation between researching parties, industrial parties and governmental organisations in projects like DURP Ondergronden could be a good opportunity to develop a vario-scale IMGeo as described in this report.

map generalisation
IMGeo
top10nl
constrained tGAP

http://resolver.tudelft.nl/uuid:d650352b-78fb-4919-9cb0-f2f66b17aef5

TU Delft, Aerospace Engineering, Geo-Database Management Center

Student theses

master thesis

(c) 2008 A.M. Hofman