Onion Leaf Disease Classification using Machine Learning Techniques

The onion, scientifically known as Allium cepa, is a fundamental food for agricultural labourers worldwide as well as people worldwide. Conversely, onion crops are susceptible to many illnesses that can greatly diminish their production. This article proposes a novel method for early detection of illnesses that damage onion leaves, utilising state-of-the-art machine learning techniques. Our collection comprises high-resolution images of onion leaves afflicted with various diseases such as downy mildew, bacterial blight, and purple blotch. By employing advanced convolutional neural networks (CNNs) such as ResNet50 and VGG16, which have been trained using transfer learning, our model can accurately detect diseases by extracting and categorizing visual input. Ensemble learning is more efficacious than data augmentation in enhancing predictions. Data augmentation expands the training dataset to enhance adaptability. In addition, the traditional methods such as support vector machines (SVM), decision trees, and random forests with convolutional neural networks (CNNs) are used. The results of the comparison show that the SVM algorithm is the most efficient and precise technique for diagnosing illnesses in onions leaves. Several commonly used metrics for evaluating model performance include the F1 score, recall, accuracy, and precision. The proposed method allows for non-intrusive continuous monitoring, empowering farmers to proactively mitigate the impact of diseases on onion harvests. This work contributes to the field of precision agriculture by applying machine learning techniques to enhance crop management in an effective and sustainable manner. This research study offers a comprehensive framework for promptly and accurately detecting onion leaf diseases, which achieves an essential need in the agricultural industry.