Deep Learning for Agriculture: Sweet Corn Disease Detection with Convolutional Neural Networks

Abstract

Sweet corn crops have a significant detrimental impact on agricultural productivity and food security because of their high vulnerability to several diseases. A timely and accurate illness diagnosis is essential for both disease treatment and yield enhancement. This paper suggests an automated method for classifying and identifying sweet corn diseases from leaf pictures that is based on convolutional neural networks (CNNs). The collection, which is separated into three categories—common rust, northern corn leaf blight, and maize dwarf mosaic virus—includes images of both healthy and damaged sweet corn leaves. Preprocessing techniques including scaling, normalization, and augmentation are applied to the images in order to increase the model's durability. The CNN design uses many convolutional layers, batch normalization, and dropout to maximize feature extraction and reduce overfitting. The categorical cross-entropy loss function and the Adam optimizer are used to train the model in order to achieve high classification accuracy. Performance is evaluated using precision, recall, F1-score, and confusion matrix analysis. Experimental results show that the proposed CNN model outperforms both pre-trained architectures and traditional machine learning methods in the classification of sweet corn illnesses. By facilitating proactive crop management and real-time disease identification, this work represents a major step toward incorporating AI-driven solutions in smart agriculture. Future studies will concentrate on developing a mobile application for field deployment in the actual world, adding attention methods, and growing the dataset.