Machine Learning for Strategic Urban Planning

Data mining is an important part of strategic planning for the development of modern urban settlement with capacities to accommodate population explosion. Developing countries are fast becoming urbanized giving the developments and opportunities that are lacking in rural areas. Data regarding urban development such as satellite image need to be analysed to ascertain the possibilities for further development or opening up of new settlements. This work presents a binary sub-pixel and feature based method of classification to detect water bodies and vegetation in earth observatory images. In this work, the images were subjected data pre-processing, feature extraction, and analysed the data using machine learning classification method to detection regions that support urban expansion or development of new settlement. The proposed method achieved 88.93% accuracy and 0.06% RMSE.