Artificial Intelligence Assessment of Mammalian Diversity Across Fragmented Agricultural Landscapes

Abstract

Background: Habitat fragmentation through agriculture is among the major reasons for changes in the composition and diversity of mammalian species in working landscapes. Traditional field studies of mammalian diversity are costly, limited to specific geographic locations, and insufficiently timely to provide useful insights for decisions regarding land use. Approach: This paper proposes a theoretical framework on how artificial intelligence (AI), specifically CNN-based image recognition technology, can be used in the assessment of mammalian diversity in fragmented agricultural landscapes. The framework combines the data from camera traps and remote sensing, automatic species classification, diversity index calculation, and metrics of landscape fragmentation into a single pipeline, without making any reference to any particular data source or mathematical formula. Discussion: The framework is evaluated based on the results presented in the literature on habitat fragmentation and AI-assisted wildlife monitoring, and a comparative table with the predicted responses of mammalian diversity to fragmentation is presented. Conclusion: AI-assisted monitoring provides an efficient tool for assessing mammalian diversity in fragmented agricultural landscapes.