Urban land use change detection based on RS&GIS

Urban land use change may influence natural phenomena and ecological processes. Decreasing of cultivated land area in Chengdu is one of the critical problems in recent years. The objective of this study is to detect land use changes between 1992 to 2008 using satellite images of Landsat TM/ETM+. The land use maps of 1992 waw collected from 1:50000 digital maps and Arc/View 3.2 software. TM/ETM+ satellite data were used to generate land use map. The images quality assessment and georeferencing were performed on images. Different suitable spectral transformations such as rationing, PCA, Tasseled Cap transformation and data fusion were performed on the images in ENVI4.0 and ERDAS IMAGE 8.5 software. Image classification was done using supervised classification maximum likelihood and minimum distance classifier utilizing original and synthetic bands resulted from diverse spectral transformation. The result of change detection shows that the cultivated land area decreased between 1992 and 2008 by 31.683% from 21.127×10⁴hm² to 14.433×10⁴hm², and the built-land area increased between 1992 and 2008 by 66.318% from 4.427×10⁴hm² to 7.363×10⁴hm². Also, the area with irrigated land farms have been decreased to 6.694×10⁴hm² by 75.343% and the dry land farming area increased to 2.728×10⁴hm² by 9.2%. The overlaying map of land use change shows that intensity of conversion of arable lands to built-land increased, and its speed of spatio-temporal change of urban land use is notable.