Monitoring The IIUM River Using Unmanned Aerial Vehicle And Image Classification

Abstract

Abstract Prior research has shown viable methods towards identifying sources of pollution in rivers by utilizing Unmanned Aerial Vehicles (UAVs) combined with proper image classification techniques. This research attempts to develop and implement a novel approach to monitor the IIUM River whereby a Parrot Bebop 2 drone is utilized for data collection, while the Quantum Geographic Information System (QGIS) software is used for the supervised classification of the collected data. The image processing techniques of stitching or mosaicking, georeferencing and supervised classification are done using Adobe Photoshop, QGIS Georeferencing plugin, and QGIS Semi-Automatic Supervised Classification Toolbox, respectively. Results show that the classification process successfully recognized target objects, however, differing sun locations in datasets along with insufficient training data have led to some minor flaws. Despite these flaws, this research successfully achieved its objectives and will be vital for further investigations in the future.