Experimental study of urban growth pattern classification using moving window algorithm

Abstract

Urban growth pattern can be generally categorized as either infill, expansion or outlying growth. Moving window algorithm determines urban growth pattern based on moving window analysis and a set of classification rules. However, literatures are concerned that the existing algorithm may produce incorrect classification result as it is strongly influenced by the size of moving window frame and classification rule. This study aims to investigate the effect of different moving window frames on the classification results and proposed an improvement to moving window algorithm with new classification rules. Results show that the existing algorithm is only able to classify outlying growth whereas the improved algorithm is not only able to classify outlying growth, it can also classify infill growth. © Medwell Journals, 2016.