Machine Learning in Spectral Imaging for Smart Farming: A Review

This article takes a detailed look at how Machine Learning (ML) is applied on spectral imaging to improve smart agriculture. In a context where agriculture faces crucial environmental challenges, spectral images become fundamental tools to understand the conditions of crops and their environment. The review focuses on evaluating how various ML techniques address challenges such as high dimensional and noise in these images. By following strict inclusion and exclusion criteria, recent and significant studies in key areas such as crop stress detection and optimization of agricultural practices are highlighted. We summarize key contributions and technological advances, providing a comparative analysis of the techniques used in reviewed studies and highlight the effective convergence of ML and spectroscopy as an essential driver for agricultural efficiency and resilience.