engleski

Urban growth pattern detection and analysis

This research presents an overview of the existing remote sensing methodology for data collection for detection and analysis of urban growth pattern. After the introduction, the topic of selecting optimal satellite images for the purpose of growth pattern detection was presented. The research deals with the satellite image preprocessing, classification of satellite images, and methods for the land- cover map optimization, i.e., postprocessing. Furthermore, some methods (and measures) for the accuracy assessment of the land-cover classification are discussed. The entire process of growth pattern detection and analysis in urban areas was shown for the city of Zagreb, Croatia. The research study period was 35 years. The change detection analysis was made by comparing the land-cover maps made with the SVM (Support Vector Machine) supervised classification method based on the 30 m bands of Landsat-5 and Landsat-8 satellite imagery. The analysis showed large urbanization in the city of Zagreb in the last 35 years. For the most part, urbanization was based on the change of barren land or low vegetation classes into the built-up area. The results show that urbanization is accompanied by an increase in population. It should also be noted that open- source data and open-source software were exclusively used to conduct this research. Therefore, presented procedures can be easily and free of charge used in the detection analysis of urban growth pattern for other urban areas in the world.

remote sensing ; growth pattern ; urbanization ; classification ; Landsat

nije evidentirano