ABSTRACT
This research investigates the use of Internet of Things and machine learning technologies in enhancing precision agriculture, having a special focus on crop management, for instance, rice cultivation. In this respect, an IoT-based system is designed, with the use of temperature, humidity, water level, and soil moisture sensors that allow real-time monitoring of the state of the crops. Then, various machine learning models are developed, such as Artificial Neural Networks , Support Vector Machines , Long Short-Term Memory networks, and Decision Trees . The whole dataset comprises 2800 sensor readings that were collected during a month are used to train models and test their performance. In particular, different models can produce diverse results, with some of them leading to higher accuracy in predicting rice water requirements. For example, the best result of 98.10% is found for ANN, followed by 94.5% for LSTM, 92.3% for SVM, and 90.25% for DT. Moreover, the trained model of ANN can be used to provide real-time prediction of rice water needs after being supplemented by a sensor that will transmit the data to the cloud. The transmitted data can then be visualized panel in the application intended to help farmers in making decisions . Overall, the research demonstrates the efficiency of IoT-based machine learning systems in improving irrigation and enhancing crop yield to ensure sustainable agriculture. In the future, the current research can be expanded by replacing the use of other crops or agriculture in general and considering the employment of more complex algorithms.
