A Hybrid Deep Learning Framework for Automated Detection and Classification of Maize Leaf Diseases

The classification of plant leaf diseases is a critical step in safeguarding global food security and ensuring sustainable agricultural productivity.Recent advances in artificial intelligence (AI) and deep learning have accelerated the development of automated disease detection systems, improving both efficiency and diagnostic accuracy.This study introduces a novel hybrid strategy that integrates Convolutional Neural Networks (CNN) with traditional machine learning classifiers, specifically Support Vector Machines (SVM) and XGBoost, to leverage the complementary strengths of both paradigms.Two hybrid variants, CNN+SVM and CNN+XGBoost, are proposed, alongside the use of Particle Swarm Optimization (PSO) for feature selection, enabling the extraction of the most informative features for robust classification.The dataset used for experimentation comprises 15,000 maize leaf images, covering four major diseases: Northern Corn Leaf Blight, Common Rust, Brown Spot, and Gray Leaf Spot, along with healthy samples.Experimental evaluations confirm that the hybrid models outperform standalone CNN classifiers, achieving superior accuracy and generalization.Specifically, CNN+XGBoost achieved 95.2% accuracy, 93% precision, 96.3% recall, and 94.3% F1-score, while CNN+SVM achieved 92.7% accuracy, 91.1% precision, 94% recall, and 92.2% F1-score.These results demonstrate the synergistic benefits of combining CNNs with classical classifiers, while PSO significantly enhanced classification outcomes by identifying highly discriminative features.This research contributes to the advancement of precision agriculture by presenting a scalable hybrid framework for maize disease detection, offering timely diagnostic insights for farmers and promoting sustainable crop management practices.