Transitioning to a Sustainable Smart Agriculture Using Deep Machine Learning Techniques: The Case of a Greek Project

In these times of global crisis, agriculture has been confronted with many challenges that must be met in order to serve a rapidly expanding global population and contribute to the sustainability of the planet. Hence, Sustainable Smart Agriculture could provide better and faster solutions to the challenges that need to be addressed. The concept of smart agriculture refers to the use of modern information and communication technologies in agricultural production. The purpose of this chapter is to demonstrate how to transition to a smart, sustainable agriculture by using deep machine learning techniques. Additionally this chapter discusses a research effort and an initiative developed by International Hellenic University, a Greek university located in a rural area. A web service based on deep machine learning has been developed by the research team to monitor crops and promote agricultural sustainability. Using drone images of various crops at different developmental stages, models were trained to diagnose and treat affected plants as quickly as possible. Plant health is improved while avoiding the use of pesticides and fertilizers that cause numerous environmental issues, such as eutrophication and water pollution, as well as posing a health threat to humans. The web-based application enables Greek farmers to upload a picture of their crops and receive a prediction for decision-making. Consequently, agricultural yields are expected to increase, while transitioning to sustainable smart agriculture is expected to be promoted.