Determination of rainwater harvesting potential in GIS using UAV imagery with machine learning classification

Abstract

The importance of water is increasing with the growing world population. Therefore, preserving and collecting water is essential for the continuation of life. Rainwater harvesting is one of the long-standing methods of water collection. With advances in Geographic Information Systems (GIS), Unmanned Aerial Vehicle (UAV) based remote sensing and Machine Learning (ML) image classification technologies, it has become easier to calculate rainwater harvesting potential. However, methods need to be customized for different man-made objects. In this study, the rainwater harvesting potential of a building complex with its surrounding marble ground surface located in the Harran University Osmanbey Campus was determined in GIS by using UAV imagery with ML classification techniques. The results showed that a 4153 m² grass area can be irrigated every day for a year from this potentially harvested rainwater.