Analysis of urban morphology through spectral mixture analysis (SMA) technique in remote sensing

Abstract

Urban morphology is the phvsical appearance ofsocial reality. Aualvzing of urban morphology is necessarvfor predicting the urban changes such as climatic change and for understanding the urban processes such as urban sprawl 01110 agricultural lands & wild habitats, and the expansion of residential land use at the urban [ringe Earlier, aerial or field sutvevs lI'ere commonly used/or producing land use //laps in order to understand the spatial structure of the cities. However. these manual methods are more labor intensive. costlv and difficult to defect morphological part ems over a long period of till/e. Thus this research attempted to find out the feasibility of using nell' technologiesfor urban morphological aualvsis. The technique adopted in this research 1ms Spectral Mixture Analysis (SA1A), which is one 0/ the techniques in Remote Sensing technology. Two IKONOS sate/lite images captured in ]005 covering Colombo district lI'ere used in this research. The commercial image processing software. ENVI4.1 was used for data analysis & visualization. This work evaluates the morphological characteristics 0/ Dehiwala-Mount Lavinia Municipality. using end-members that 1I'ere extracted through Pixel Purity Index (PPI) , Scatter Plots & n-Dimensional visualizer. The abundance maps 1I'ere generated for impervious surfaces, green vegetation, bare soil & shade. Classification results [rom applying the SMA 11'ere assessed by comparison with ground reference data. The study identified tliat SMA technique classified the image with a high accuracv gil'lllg 1I'ind011's to identify major land use types and urban morphology with a considerable accuracy,