Urban expansion analysis based on spatial variables derived from multi-temporal remote sensing imagery

In this research, we focus on the spatial pattern of the urban expansion. The spatial pattern of the urban area can be quantitatively delineated by many spatial variables. Numerous spatial variables have been examined to evaluate their applicability to the urban change. These metrics include road network accessibility, built-up density and some landscape metrics. Remote sensing technology was used for monitoring dynamic urban change. Multi-temporal Landsat TM images (1988, 1991, 1994, 1997, 2000, and 2002) were used for the change detection using post-classification comparison method. The road network and its change were extracted from multitemporal images using the GDPA algorithm. Contagion, one of the landscape metrics, was selected, because it it can describe the heterogeneity of the suburban area, where the landuse change is most likely to happen. Analysis has also been conducted to identify the relationship between urban change and these spatial variables.