Optimal Routing and Deep Regression Neural Network for Rice Leaf Disease Prediction in IoT

Abstract

To meet the increasing food demand, production of rice is increased. Unfortunately, rice leaf disease has caused a major problem in the agricultural yield. Various disease prediction strategies are developed in the Internet of Things (IoT) agricultural applications, but accurately predicting the disease causes substantial environmental issues. Therefore, an effective method named Sunflower EarthWorm (S-EWA) optimization algorithm is proposed in this research to predict the disease in the rice crop. The sensor nodes are dispersed randomly in the IoT network of agricultural field, and these sensor nodes collect the agricultural data from the rice crop and are sent to the base station (BS) through the optimal path, which is computed using the proposed Sunflower EarthWorm optimization algorithm. The regeneration, the reproduction, and the dynamic behavior of the optimization algorithm effectively transfer the data through routing using the optimal path. The optimization algorithm uses the fitness function to determine the optimal path based on the position of earthworms. Deep Regression Neural Network uses the artificial neurons and performs the disease prediction of rice leaf at BS. The proposed S-EWA- based DBN attained better performance in terms of accuracy as 95.2, sensitivity as 95.51, and specificity as 94.89 by varying the training percentage, and accuracy as 95.7, sensitivity as 95.86, and specificity as 95.54 by varying the hidden layers, respectively.