Building-based urban land use classification from vector databases in Manchester, UK

Abstract

The recognition, analysis and classification of urban structures are important in urban land use modeling. The form and the function of individual urban elements such as buildings and street blocks help us better understand the urban morphology. The types, layout and arrangement of these buildings form up the local characteristics of urban areas. A model has been developed to classify urban areas based on the cartometric properties of buildings and the patterns they make. Supervised and un-supervised classification algorithms from data mining techniques along with GIS are explored to help create a framework for extracting information from vector databases and classifying building and blocks. The methodology is developed and applied to Manchester metropolitan in the UK.