Contents Abstract i Samenvatting v Acknowledgements ix Contents xi 1 Introduction 1 1.1 Background and motivation 1 1.2 Problem definition 7 1.3 Research objectives 10 1.4 Methodology 14 1.5 Thesis outline 16 2 Literature review 19 2.1Patterns and structures in the geospatial domain 20 2.2 Evaluation of generalization quality 25 2.3 Automated evaluation of generalization output 31 2.4 EuroSDR project 44 2.5 Summary 48 3 Conceptual framework 49 3.1Framework overview 50 3.2 Analysis of map specifications 52 3.3 Evaluation of cartographic constraints60 3.4 Balanced map specifications 63 3.5 A three-step approach to automated evaluation 66 3.6 Summary 68 4 A formal theory for automated evaluation 69 4.1 Spatial entities and relationships 70 4.2Characterization and evaluation functions 83 4.3Constraints as logical expressions 85 4.4Discussion and reflections87 5Design of data enrichment 93 5.1 Introduction 93 5.2Data enrichment overview 94 5.3Patterns in building clusters 101 5.4Recognition of building alignments 108 5.5Measures of spatial distribution 120 5.6Similarity between spatial distributions 127 6Design of data matching and evaluation 137 6.1 Introduction 137 6.2Data matching using pattern classification 139 6.3Matching of building patterns 148 6.4Hvaluation of building alignments156 6.5Evaluation of spatial distribution 164 7Implementation and validation 169 7.1Introduction 169 7.2Building clustering 170 7.3Recognition of building alignments172 7.4Data matching via supervised classification 183 7.5Preservation of building alignments 191 7.6Preservation of spatial distribution 200 8Conclusions and recommendations 211 8.1Conclusions 212 8.2Contributions225 8.3Recommendations for future work 229 A Auxiliary geometric structures 233 A.1 Delaunay triangulations 233 A.2 A Voronoi-like diagram based on skeleton 237 B Expert survey 243 Bibliography 251 Author's Biography 271