A review of Smart Irrigation

Providing reliable access to water is one of the critical challenges of this century. Applying the exact amount of water over the crops may substantially increase yields using fewer resources and reduce the environmental footprint. Current mechanistic models guide the irrigation process by describing how water transitions between soil, plants, and the atmosphere. New approaches take advantage of the recent convergence of technologies (Internet of Things, Artificial Intelligence, Big Data, optimisation methods) to transform agricultural measurements into accurate estimations of the crop's water needs while considering practical constraints. For global agriculture, scientific and technological advances are converging toward fully automated irrigation management. However, there are still complex challenges to overcome before this prospect comes true, starting by persuading farmers of its effectiveness. This paper provides an overview of past and current irrigation practices and future directions toward Smart Irrigation. It aims to offer a reference for state-of-the-art Smart Irrigation and outline future possibilities. Starting with a staged view of the irrigation evolution, the existing literature is reviewed in terms of problem and solution aspects, including a discussion on prominent issues to overcome. This review identifies data-driven and machine learning models as key enablers for farmer-friendly – and eventually farmerless – irrigation, while highlighting ongoing challenges in sensor deployment, farmer adoption, and system interoperability.