An IoT based Tree Specific Soil Nutrient Management System Using Neural Network for Cashew Cultivation

Abstract

Anacardium occidentale commonly known as Cashew was introduced into India by Portuguese in 16^th century. Farmers of Cuddalore District, Tamil Nadu in India, are interested in Cashew Cultivation due to its export market values. To achieve better yield from cashew cultivation, farmers need to supply balanced nutrition to the soil and by following intercropping method. In literature, Recent Technologies like Artificial Intelligence and IOT are widely used to predict dosage of the fertilizer to be applied. By examining the numerous connected ascribe success locations of Cashew orchards, it was possible to determine how much macronutrients like Nitrogen (N), Phosphorus (P), and Potassium (K) were present in the soil and it will be helpful for the understanding of soil fertility level of the area. In our work, using IOT set up which includes Arduino UNO, NPK sensor and OLED display, we identified NPK values of soil for various cashew trees in a particular area. Along with NPK values of soil, inputs include the local soil type, PH and Tree age. Based on the input factors, DL trained model will provide suggestions of fertilizer dose to be applied for a specific tree for the Cashew cultivating farmers for better yield.  In our proposed work, LSTM based Recurrent Neural Network (RNN) Algorithm is used to provide a better prediction of fertilizer. With the help of this research, we can be provided with what intercrop to plant and how much fertilizer to use, in what ratio to supply and fertilizer shop suggestion to enhance agricultural knowledge across the globe using the Internet of Things (IoT) and Artificial Intelligence. The result of RNN model was compared with Random Forest, Linear Regression, Linear Random Classifier, Decision Tree Classifier with proving accuracy of 98.7 %. Together with a decrease in the farmers' input efforts, these strategies will increase the productivity of the fields.