Developing a Hybrid Faster Recurrent Convolutional Neural Network with an Improved Weighted Artificial Bee Optimization Method for Grape Disease Detection at an Early Stage

Abstract

In India’s diverse range of crops, fruits play a crucial role in generating significant revenue for farmers. Among these fruits, grapes are extensively grown. However, grape plants are susceptible to various diseases affecting their fruits, stems, and leaves, ultimately impacting their yield. Grape disease detection at an early stage plays a crucial role in ensuring crop health and maximizing yield. To address this issue, early detection and effective treatment of these diseases are essential to ensure food safety. This study focuses on analyzing different methods for diagnosing and classifying diseases that affect grapevines, with particular emphasis on grape leaf diseases. Monitoring the condition of grape leaves provides valuable insights into the overall health of the grape plants. The research aims to provide a comprehensive overview of techniques used for identifying and categorizing these diseases. Automated disease detection algorithms are proposed to enhance diagnosis accuracy and enable timely control actions. Image processing, a widely used method, is endorsed for leaf disease identification and classification in plants. In this study, we propose a novel approach by integrating a hybrid Faster Recurrent Convolutional Neural Network (FRCNN) with an improved Weighted Artificial Bee Optimization (WABO) method for efficient grape disease detection. The FRCNN architecture combines the strengths of recurrent and convolutional neural networks to capture both spatial and temporal information from grape images. Additionally, the WABO method enhances the training process by optimizing the network’s weights to improve detection accuracy. Experiments conducted on a large-scale dataset show that the proposed approach is more effective than current approaches in terms of computational efficiency and detection accuracy when it comes to correctly diagnosing grape diseases early on. The proposed framework has a lot of potential for practical uses in vineyard management and precision agriculture.