UAVs, Hyperspectral Remote Sensing, and Machine Learning Revolutionizing Reef Monitoring

Abstract

Recent advances in unmanned aerial system (UAS) sensed imagery, sensor quality/size, and geospatial image processing can enable UASs to rapidly and continually monitor coral reefs, to determine the type of coral and signs of coral bleaching. This paper describes an unmanned aerial vehicle (UAV) remote sensing methodology to increase the efficiency and accuracy of existing surveillance practices. The methodology uses a UAV integrated with advanced digital hyperspectral, ultra HD colour (RGB) sensors, and machine learning algorithms. This paper describes the combination of airborne RGB and hyperspectral imagery with in-water survey data of several types in-water survey of coral under diverse levels of bleaching. The paper also describes the technology used, the sensors, the UAS, the flight operations, the processing workflow of the datasets, the methods for combining multiple airborne and in-water datasets, and finally presents relevant results of material classification. The development of the methodology for the collection and analysis of airborne hyperspectral and RGB imagery would provide coral reef researchers, other scientists, and UAV practitioners with reliable data collection protocols and faster processing techniques to achieve remote sensing objectives.