Internet of things based machine learning supported selective irrigation system design and implementation

In recent years, the increase in water consumption due to rapid population growth, the decrease in natural water resources due to climate change and the unconscious agricultural irrigation result in an increase in competition between countries for fresh water resources. The world population is increasing rapidly, but in parallel with this increase, agricultural production does not increase in the same direction. One of the reasons for the inadequacy in agricultural production is the widespread use of irrigation with traditional methods. In traditional agricultural irrigation, more water is given than the plant needs and most of the water evaporates without being used by the plant. In agricultural production, the need for smart technologies in agricultural irrigation is increasing day by day in order to keep water use at the most appropriate level, reduce energy consumption and increase the quality of crops. In this thesis study; The use of water in agricultural irrigation is made efficient. An internet-based and machine-learning system design was carried out to create a plant-specific irrigation program. Soil moisture data was collected with a Wireless sensor network (WSN) based system. To assist the decision-making mechanism of the Selective Irrigation System (SSS), irrigation water data generated specifically for plants by the FAO CROPWAT program were used. In line with the data obtained, the amount of plant-specific irrigation water was estimated by applying regression algorithms. The data obtained by machine learning was sent to the NodeMCU processor with the MQTT protocol, allowing the control of the Solenoid Valves that allow the passage of water. Drip irrigation method will be used in the system. The energy requirement of our study was obtained from the Photovoltaic (FV) system.