An Automated Neural Network Approach for Predicting Paddy Leaf Diseases and Autopumping of Fertilizer along with IoT

Agriculture is an important sector that supports both human life and the economy. It entails growing crops and rearing animals for food, fiber, fuel, and other purposes. Also, crop production must be adequate for this to occur. However, attaining a correct yield of crops these days is a difficult task because they are infected by several diseases throughout their growth and often need to be recognized by the farmer, resulting in an inappropriate yield of crops. Agriculture has been practiced for thousands of years, evolving to integrate new technology and methods that improve productivity, sustainability, and efficiency. This work focused on four types of paddy illnesses and one healthy leaf class. The primary objective of this work is to identify a suitable deep learning model with higher accuracy than present deep learning models, the accuracy of which is also highly debated. We investigated three Convolutional Neural Network (CNN) algorithms, AlexNet, ResNet, and Proposed Algorithm for predicting paddy leaf diseases, which achieved an accuracy of 92%, 86.89% and 96.93% respectively. Furthermore, hardware implementation is carried out to auto-spray the suitable fertilizer for the specific predicted disease by the CNN algorithm with higher accuracy in association with Internet of Things (IoT) connectivity.