Harnessing Machine Learning for Accurate Corn Leaf Disease Classification

Corn, a vital global agricultural commodity, confronts significant threats from diseases that adversely impact yield and quality. Ensuring its stable and sustainable production is important for global food security, farmer prosperity, and national economic stability. Given the diverse challenges faced by corn cultivation, this research investigates the application of various machine learning models in the context of corn disease classification. This research provides an understanding of how various machine learning models react to various feature extraction techniques. The study compares the performance of the Histogram of Oriented Gradient (HOG), Discrete Wavelet Transform (DWT), and ResNet50 for feature extraction. The methodology encompasses image collection, preprocessing, feature extraction, and classification employing machine learning algorithms. Comparative analyses involving ResNet50, HOG, and DWT feature extraction techniques revealed that the ResNet50-based model achieved the highest accuracy. Furthermore, the results underscore the superior performance of the Support Vector Classifier (SVC) when coupled with ResNet50, highlighting its suitability for corn disease classification. This methodology not only offers a robust solution for efficient corn disease diagnosis but also contributes significantly to the advancement of agricultural disease management.