Utilizing Modified Metaheuristic Optimizers for Computer Vision Optimization in Agriculture

A strong agricultural sector is crucial for supporting global population growth. While monoculture improves production efficiency, it also increases the risk of pathogen outbreaks if not addressed promptly. Excessive use of pesticides and fungicides can reduce yields, degrade soil quality, and promote pest resistance. Early detection and effective intervention are vital to preventing crop loss and meeting growing food demands. This study aims to provide a valuable tool for the early detection of plant disease detection leveraging artificial intelligence (AI). This work presents a modified metaheuristic optimization method tailored to enhance convolutional neural network (CNN) performance for plant disease detection. A comparative analysis is conducted with various modern optimizers on publicly available datasets. The best-performing models achieve an accuracy of 75.43%, indicating potential for practical application.