Leaf Disease Detection of Medicinal Plants Using Convolutional Neural Network

Plants are a very important asset of our nature. Besides providing flowers, fruits, and vegetables there are plants with medicinal properties also which are widely used in medical and food industries. The diseases that harm these plants must therefore be completely understood. Detection of diseases in early stages can help prevent production losses. Manual detection and classification of diseases is very complicated and needs a specialized person with the required knowledge. Therefore, an automatic plant disease detection system for medicinal plants is suggested in this research paper. To achieve this CNN (Convolutional Neural Network), a deep learning-based algorithm is used to classify medicinal leaves as healthy or diseased. A high-quality leaf dataset (of 12 different plants with medicinal and economical importance) is taken and the annotated images are used for training and testing. It was observed that the feature extraction process using CNN gives 92.7% accuracy. Finally, a GUI was developed for easy application.