Smart Implementation of Plant Disease Detection System using Machine Learning

This plant disease detection system uses machine learning and computer vision to help farmers identify plant diseases early and accurately. By analyzing plant photos, farmers can immediately take preventive action against possible crop loss. The process also reduces dependence on poisonous pesticides, thus promoting healthier and more sustainable farming practices. The system is designed using a data set of more than 12,000 images of healthy and diseased plants. The images are processed to improve visibility, and the MobileNet V2 model is trained to identify disease symptoms, such as color shift, spot appearance, or wilting leaves. Farmers can use a simple mobile or web application to upload plant photos. The system then scans the image, marks the infecte d areas, and makes predictions about possible diseases, thus allowing farmers to address the problem immediately. The system is updated and improved periodically based on user feedback to ensure accuracy and adapt to new diseases or environmental changes. It has been proved to be high in effectiveness, with high accuracy rates: 88.86% in training, 94.44% in validation, and 94.19% in testing, along with precision of 93%, recall of 89%, and F1 score of 90%. Although specific diseases with similar appearances are misclassified, the system overall remains effective. With additional features such as camera capture in real- time, upload of photos, and preventive recommendations, this tool offers much support for modern farming practices.