Identification of Paddy Leaf Diseases Using a Supervised Neural Network

Abstract

Rice is the most often eaten staple food by billions of people worldwide. However, paddy crops are susceptible to diseases such as brown spot, hispa, leaf smut, leaf blast, bacterial leaf blight, and cotton mold. The purpose of this article is to identify paddy leaf diseases. The proposed scheme is built using CNN with a pre-trained ResNet-50, ResNet-101, VGG-16, VGG-19, EfficientNet, Inspection-V2, and GoogleNet library. A ReLU classifier is used to enhance the accuracy and efficiency of the identification process. This model can help the farmer recognize the paddy leaf condition as a primary diagnosis, and it can also help the agriculturist verify his prediction by examining the paddy leaf. Traditional laboratory procedures are both costly and time-consuming. This paper’s proposed paddy lead disease detection system will identify and diagnose five classes of paddy leaf conditions. The system’s highest average accuracy for all five identifications is 96.27 percent for the ResNet-50 pre-trained library, and the F1 Score is 98.19 percent. GoogleNet has the lowest classification accuracy of 86 percent. The accuracy can be maximized by expanding the dataset. This technology will take less time and less money to diagnose paddy leaf conditions than traditional procedures, and it will aim to maximize the production of paddy.