Identifying Medicinal Plants and Their Fungal Diseases

Today, with the development of technology, most manual methods are replaced by automated computer systems for the easiness of human beings. Plant identification and disease classification are two major agricultural research areas, focusing on introducing computerized systems rather than manual methods. Many researchers used various identification and classification techniques using computer-based systems as human classification errors lead to risk and high cost. Medicinal plant identification needs an expert to correctly identify plants because misidentifying poisonous plants as medicinal plants causes fatal cases. Further, taking diseased medicinal plants to prepare medicines and herbal products may have adverse effects. Therefore, this study proposed a computerized method to identify medicinal plants and classify their diseases to overcome such shortcomings. In this work, a comparison is done with Convolutional Neural Network (CNN) architecture from scratch and Transfer Learning with several experiments. Transfer learning models achieved higher accuracy than CNN architectures for medicinal plant identification with 99.5 % accuracy and medicinal plant disease classification with 90% accuracy, respectively.