Harnessing Artificial Intelligence and Remote Sensing for Large-Scale Marine Habitat Monitoring and Conservation

Abstract

Human activities and climate change are posing a growing threat to marine eco systems, which makes it necessary to have an effective and big scale monitoring solution. Field-based surveys particularly traditional ones are useful, but may be restricted by cost, time and space. This paper suggests a new method to address these restrictions using remote sensing tools and artificial intelligence (AI) to scale-up monitoring of marine ecosystems. With satellite, aerial and drone photography, machine learning (ML) and deep learning algorithms, we categorize marine habitat types, identify environmental alterations and the threats to conservation. The multispectral and hyperspectral imagery data collected using remote sensing can cover coastal and marine ecosystems in a comprehensive way, whereas AI-based methods enable automated data analysis, which can be used to map the habitat efficiently and detect changes in time. Field data, such as in-situ survey and ground-truting of the habitat maps, are used to validate the derived habitat maps and to identify the environmental change. The findings demonstrate good classification capabilities, whereby AI-based models correctly classify habitat types and detect avoidable environmental variations, including coral reef degradation and habitat loss. The results demonstrate the possibilities of this combined method in the case of large-scale marine habitat monitoring, which is a cost-effective and scalable instrument of conservation management. This will assist in the proactive conservation policy, improve policy-making, and present practical knowledge on sustainable marine resource management.