REFINING THE DETECTION OF MEDICINAL PLANTS USING MACHINE LEARNING AND IMAGE ANALYSIS TECHNIQUES (KEY IJP************822)

• Kokku Venkata Ganesh,Alladi Ramesh,Sangem Naresh,Potharaju Ajay,Dyavari Paramesh

Abstract

India boasts a rich heritage of floral diversity, a vital source of medicinal plants integral to Ayurvedic Pharmaceutics. However,the misidentification of these plants remains a critical issue, with numerous crude drugs sold under the same name in the market. This confusion stems from factors such as seasonal and geographical variations and similar morphological characteristics. Consequently, the demand for these resources has led to issues of impurity, substitution, and doubtfulness regarding their curative potential. In response, this project presents an innovative solution.It employs Image Processing and various Machine Learning Algorithms to develop software capable of identifying different medicinal plants and raw materials and listing out their uses. By doing so, it offers a transformative tool for stakeholders throughout the supply chain, from wholesalers to distributors, ensuring the authenticity and quality of Ayurvedic ingredients.This technology not only safeguards India's botanical heritage but also reinforces trust in the efficacy of traditional herbal medicine systems. The medicinal leaf dataset consists of 30 classes. Transfer learning approach was used to initialize the parameters and pretrain Neural networks namely MobileNetV2, InceptionV3, and ResNet50.These component models were used to extract features from the input images and the SoftMax layer connected to the Dense Layer was used as the classifier to train the models on the concerned dataset.