文章基本信息

标题：AN EFFICIENT CLUSTERING TECHNIQUES FOR URBAN AREA ANALYSIS BASED ON SATELLITE IMAGES
本地全文：下载
作者：A. TOBAL ; H. FAROUK ; S. MOKHTAR 等
期刊名称：Journal of Theoretical and Applied Information Technology
印刷版ISSN：1992-8645
电子版ISSN：1817-3195
出版年度：2020
卷号：98
期号：8
页码：1238-1246
出版社：Journal of Theoretical and Applied
摘要：Urban analysis provides the urban planners with the information about how to optimally utilize the land resources and the infrastructure. It gives solutions to many problems such as dense population, population growth, and limited resources. The Arab Republic of Egypt is one of the most densely populated countries in the region. In addition to the accumulation of the population in a specific area of the land, there is limited natural resources that directly affects the nature of life, agriculture and the population spreading. So, it is vital to study the changes in the geography of the land in order for the urban planners to set short- and long-term strategies to improve the quality of life and set a sustainable development plan. The Suez region was selected for such study. In this work, satellite images have been categorized into four segments: Desert, agriculture, residential and water using three clustering techniques to study the increment and decrement of urbanization, water resources, agricultural patch and dissertation. The three clustering techniques are; Fuzzy, Kohonen Neural Network and k-means. Each technique was applied on a low-quality resolution Google satellite images of Suez area across 16 years from 2001 to 2016. A comparison between each technique behavior on this image style and a ready-made program ArcGIS has been done. Astoundingly, the results show that the Fuzzy clustering is the best technique for such kind of images.
关键词：Clustering Techniques;Unsupervised Learning;Fuzzy Clustering;K-means Clustering;Kohonen Neural Network.