Feature Extraction With Optimization Model In Plant Disease Detection Using Rule Generation Algorithm And Machine Learning Technique

Abstract

Plant diseases pose a serious threat to global food security, and timely detection remains challenging due to the limitations of conventional diagnostic methods, which are often slow, require expert intervention, and lack sufficient accuracy for practical applications. To address these challenges, this study proposes an enhanced image-based detection framework that improves both precision and reliability in identifying plant diseases. The framework integrates Enhanced Feature Fractal Fusion (EFT) for robust feature extraction with a hybrid optimization-driven classification strategy combining Particle Swarm Optimization (PSO) and Levenberg-Marquardt (LM) algorithms. This combination enables effective recognition of complex disease patterns from plant leaf images under diverse conditions. Experimental evaluations demonstrate superior performance, achieving 99% accuracy, 98.99% precision, 99.71% recall, and 99.99% F1-score, outperforming existing state-of-the-art methods. By facilitating rapid and accurate disease identification, the proposed approach supports early intervention, efficient crop management, and sustainable agricultural productivity, making it highly suitable for real-world precision agriculture applications.