ABSTRACT
The study focuses on using machine learning techniques to diagnose diseases affecting wheat crops, with an emphasis on common diseases such as brown spot, powdery mildew, and Fusarium head blight. A dataset of 1,430 images of healthy and diseased wheat plants was created to train the machine learning models. The study compares the performance of Convolutional Neural Network (CNN), VGG-16, K-Nearest Neighbors (KNN), and Recurrent Neighbors (RNN) in distinguishing between healthy and diseased plants. The CNN model achieved the highest accuracy at 97.77%, followed by VGG-16 with 95.5% accuracy. The results demonstrate the potential of machine learning models in effectively identifying and diagnosing crop diseases. By integrating technology-driven image extraction methods with machine learning algorithms, the study contributes to advancing crop management practices. The research has significance in enhancing agricultural productivity and ensuring food security in the global food chain.
