Observing mega cities to grow from space and predicting its growth using machine learning techniques

Abstract

In recent years, developing countries are undergoing a massive change in growth in the urbanization. Urbanization is yielding many positive effects for these developing countries in infrastructure, economy, social status and so on. However, this extensive urbanization mostly resulting some adverse effects. This study aims to evaluate and observe the various changes like vegetation, built-up and water in the urban area of the greater Dhaka area of Bangladesh which has one of the highest growth rate among developing countries using Landsat 7 ETM+ and Landsat 8 OLI images between 2012 to 2018 using different GIS and remote sensing techniques. The analysis shows the superabundant growth of the buildup areas as well as degradation in the vegetated and water areas of greater Dhaka, Bangladesh. Based on such data, different machine learning techniques have been applied to show the performance in terms of accuracy and forecast a predicting rate for the future growth of Dhaka city. The experimental results provides an idea of the subsequent changes in terms of water body, vegetation and built-up areas for Dhaka city, with an aim to provide these useful information for the policy makers and urban planners.