Identification of Different Medicinal Plants/Raw materials through Image Processing Using Machine Learning Algorithms

Identification of Different Medicinal Plants/Raw materials through Image Processing Using Machine Learning Algorithms

1Gundala Naveen, 2T Ganesh Varma, 3S Mahammad Kalam, 4A Venkatesh Varma, 5Vipul G,

6Jayavadivel Ravi

1,2,3,4,5UG Student Dept. Of CS&E, 6Assistent Professor Dept. Of CS&E

1,2,3,4,5,6 Presidency University Bangalore,560089

1nanaveen2003m@gmail.com, 2ganeshvarma.trimulasetti@gmail.com, 3mahammadshaik435@gmail.com, 4varmakbm9@gmail.com, 5vipulgmdl@gmail.com, 6jayavadivel.ravi@presidencyuniversity.in

Abstract

India's traditional medical systems rely extensively on medicinal plants, yet accurate identification remains a persistent challenge due to morphological similarities, regional and seasonal variations, and limited botanical expertise among collectors and traders. These issues lead to adulteration and substitution, thereby diminishing treatment efficacy and eroding public trust in Ayurvedic remedies. This project presents a machine learning-based image processing system to automatically classify medicinal plants from leaf images. A Convolutional Neural Network (CNN) architecture, trained on a curated dataset using PyTorch, demonstrated an accuracy of 87%, with high precision and recall. This automated system offers a scalable solution to improve quality control across the herbal drug supply chain.

Keywords:

Medicinal Plant Identification, Image Processing, Machine Learning, Convolutional Neural Network (CNN), PyTorch, Ayurvedic Medicine, Leaf Classification, Deep Learning, Herbal Drug Authentication, Computer Vision