An AI-driven Robust Strategy for Automated Grape Leaf Disease Classification with Pesticide Recommendation System using Walrus-Optimized CNN

Department of EIE, CVR College of Engineering, Ibrahimpatnam, Telangana, India

^* Correspondingauthor: k.uday@cvr.ac.in

Abstract

Agricultural productivity is crucial for global economic development, but crop diseases can significantly impact output. Early detection is essential and so, Artificial Intelligence (AI) offers a solution by classifying leaf images as healthy or diseased types for aiding farmers. Grape leaf disease stands as the major issue behind large-scale orchard losses, impeding the healthy and sustained growth of the grape industry. This research outlines a new AI-driven model comprising data collection, pre-processing, image segmentation, disease classification, pesticide recommendation, and performance evaluation. It starts with extracting a leaf dataset, applies a Gaussian Filter for noise reduction, uses Mask R-CNN for image segmentation, and employs Walrus Optimization (WaOA)-driven Convolutional Neural Networks (CNNs) for disease classification. This model then suggests pesticides based on AI analysis and evaluates their performance. Ultimately, it aims to aid farmers by minimizing losses and enhancing production through timely disease detection and effective treatment recommendations.