Implementation irrigation system using Support Vector Machine for precision agriculture based on IoT

Abstract

The industrial revolution 4.0 is driving increased production and automation of systems using the internet of things and data mining. Climate change and young farmers knowledge make it necessary to create an internet-based and Artificial intelligent farming system. Farmers need a system that can help the process of watering and fertilising for crop growth. In several previous studies, calculating the prediction of water density from moisture sensor data and to predict future moisture values using SVM, the implementation of artificial intelligence for precision agriculture to detect field-grown cucumber, in other studies determining the amount of water using fuzzy systems and neural networks, parameters used in determining irrigation systems consist of environmental data including ambient temperature, humidity and soil moisture several additional variables in determining irrigation control using crop water demand, soil evapotranspiration and weather condos, forecasting agriculture water uses LS-SVM to improve accuracy and speed in the forecasting process implementation of automated irrigation using IoT. In this paper, we propose the integration of water quantity predictions on plants using several environment variables for forecasting using linear Support vector machine. Accuracy rate of SVM reaches 95%, 94.3% precision, 91% recall, F1-score 92.7%. The results of the forecasting will be sent to the end node using the MQTT protocol.