Using K-means algorithm classifier for urban landscapes classification in Taipei area, Taiwan

Current presentation summarizes spatial analysis studies of Taipei urban growth using ENVI GIS based image classification. The presentation consists in two parts. The first part describes the city, urban and social settings and gives a brie history of the development in 20th century. The second part is focused don the GIS based technical description of the algorithms of image analysis: classification of the multi-temporal Landsat TM series of the selected stud area of Taipei, Taiwan. Methodology aims at spatio-temporal analysis of urban dynamics in study area during 15 years (1990-2005). Research objective: application of geoinformatic tools, remote sensing data and application of methodology to spatial analysis for urban studies, a case study of Taipei. Current presentation consists in 2 parts: 1) Overview of the environmental research problem, urbanization and characteristics of Taipei. Consequences of urban sprawl for the global cities, such as Taipei; 2) Detailed technical description of the GIS part: remote sensing data capture, pre-processing, algorithm processing, image classification and spatial analysis. The spatial analysis performed by means of GIS ENVI enabled to use satellite images for social and urban studies. The spatio-temporal analysis was applied to Landsat TM images taken at 1990 and 2005. Built-in functions of the mathematical algorithms (K-means) enabled to process raster Landsat TM images and to derive information from them.