DefinePK

Abstract

Cotton is a vital agricultural crop, and its yield is significantly affected by leaf diseases. Early detection and classification of these diseases are crucial for effective disease management, improving crop health, and maximizing yield. Traditional methods for disease detection rely on manual inspection, which is often time-consuming, labor-intensive, and prone to human error. To address these limitations, this study explores machine learning and deep learning-based classification techniques for detecting cotton leaf diseases.
In this research, five machine learning models—Support Vector Machine (SVM), k-Nearest Neighbors (KNN), Naïve Bayes (NB), Decision Tree (DT), and Neural Networks (NN)—were employed to classify diseased and healthy cotton leaves. The dataset, obtained from Kaggle, consists of images labeled into four categories: Diseased Cotton Leaf, Fresh Cotton Leaf, Diseased Cotton Plant, and Fresh Cotton Plant. The images were preprocessed through resizing, normalization, and data augmentation techniques to enhance the models’ robustness and generalization ability.
The performance of each classification model was evaluated using standard metrics, including accuracy, precision, recall, F1-score, and confusion matrices. Experimental results indicate that Neural Networks and SVM achieved the highest accuracy (96%), demonstrating superior classification performance. In contrast, KNN showed the lowest accuracy (65%), likely due to its sensitivity to high-dimensional data and noise. Decision Tree and Naïve Bayes achieved moderate classification performance, each with an accuracy of 75%.
The study’s findings suggest that deep learning models, particularly Convolutional Neural Networks (CNNs), outperform traditional machine learning approaches in image-based classification tasks. Future work will focus on optimizing deep learning architectures, integrating real-time classification systems using Edge AI, and expanding the dataset with more diverse samples to improve model generalization. The implementation of automated disease classification can significantly aid farmers in early detection and timely intervention, ultimately reducing crop losses and enhancing agricultural productivity