A Spatial Structural and Statistical Approach to Building Classification of Residential Function for City-Scale Impact Assessment Studies

Authors:
Dimitrios P. Triantakonstantis;Stuart L. Barr
Affiliations:
School of Civil Engineering and Geosciences, University of Newcastle, Newcastle upon Tyne, NE1 7RU;School of Civil Engineering and Geosciences, University of Newcastle, Newcastle upon Tyne, NE1 7RU
Venue:
ICCSA '09 Proceedings of the International Conference on Computational Science and Its Applications: Part I
Year:
2009

Citing 4
Cited 0

A ROC-based reject rule for dichotomizers

Pattern Recognition Letters
Fusion of point-based postal data with IKONOS imagery

Information Fusion
Textural and local spatial statistics for the object-oriented classification of urban areas using high resolution imagery

International Journal of Remote Sensing
Linear regression with special coefficient features attained via parameterization in exponential, logistic, and multinomial-logit forms

Mathematical and Computer Modelling: An International Journal

Quantified Score

Hi-index	0.00

Visualization

Abstract

In order to implement robust climate change adaption and mitigation strategies in cities fine spatial scale information on building stock is required. However, for many cities such information is rarely available. In response, we present a methodology that allows topographic building footprints to be classified to the level of residential spatial topological-building types and corresponding period of construction. The approach developed employs spatial structure and topology to first recognise residential spatial topological types of Detached , Semi-Detached or Terrace . Thereafter, morphological and spatial metrics are employed with multinomial logistic regression to assign buildings to particular periods of construction for use within city-scale impact assessment studies. Overall the system developed performs well for the classification of residential building exemplars for the city of Manchester UK, with an overall accuracy of 83.4%, although with less satisfactory results for the Detached period of construction (76.6%) but excellent accuracies for the Semi-Detached residential buildings (93.0%).