A Comprehensive Study on the Advance Methods for Detection of Plant Disease by Using Hybrid Deep Learning Method

Abstract

The expansion and development of the world economy are significantly influenced by agricultural production. Disease-affected crops harm a country's agricultural productivity and financial resources. Most of the research works have concentrated on a particular type of plant. Hence there is a need for the development of the generic Plant Disease Detection (PDD) system. Hence, to overcome the research gap, this study presents an improved framework for plant disease detection utilizing a hybrid deep learning approach that integrates Generative Adversarial Networks, Convolutional Neural Networks (CNNs), and Recurrent Neural Networks (RNNs). Detect images of various plant species affected by fungal, bacterial, and viral disasters are collected in diverse conditions, including different lighting, environmental, and other factors. The hybrid deep learning framework for PDD effectively combines dataset augmentation and transfer learning to enhance real-time accuracy, achieving 92% precision in identifying infection types and stages. This will help detect the disease in real time, allowing farmers and other agricultural stakeholders to intervene or make informed decisions. This study significantly contributes to the development of innovative technologies in agriculture and serves as a basis for research to automate plant health monitoring, which in turn can promote sustainable agriculture and also help enhance food security around the world.