A decision-making approach for Smart Farming

Abstract

Since the population is continuously increasing, by 2050 the global population will reach 9.7 billion people. The production of the food to supply this huge amount of the population has become a big challenge for modern agriculture, it is necessary to find a faster and more efficient way for food production. Therefore, new technologies are being applied in farming constructing a new era of smart farming, and Decision Support System (DSS) is one of the most important components. In this thesis, an approach of implementing a DSS is presented and the topic is focused on crop management. And more specifically, rice is the crop under study and dataset from Chinese National Germplasm Repository is used. Different technologies are involved in implementing the DSS such as Big Data, machine learning or cloud platform. Kafka streaming system is used as a messaging system to send rice examples through a direct stream connection established between Spark and Kafka; The use of Spark streaming tool enables the data processing in a distributed way; Different machine learning algorithms are trained in order to classify the rice quality of the examples and the prediction will be done in real-time in Spark; And the cloud platform used is Onesait, where final processed data is stored and dashboard visualization is fulfilled.