HerbVision: Herbal Plant Detection

HerbVision: Herbal Plant Detection

Raghav Bochare, Priyanshu Dubey, Pranav Ojha, Pranjal Nimbodiya, Ruchir Jain

Acropolis Institute of Technology and Research

Abstract:

Herbal plants play a significant role in traditional medicine and pharmacology, but accurately identifying these species in natural environments presents ongoing challenges. This study introduces a novel approach to herbal plant detection utilizing machine learning and computer vision techniques, which significantly enhances both the speed and accuracy of species identification. A deep learning model, employing Transformer Model for feature extraction and classification, was trained on a diverse dataset of high-resolution images of herbal plants. Our model achieved an accuracy of 99% in field tests, successfully addressing environmental challenges such as variations in lighting, angle, and occlusion. Optimized for real-time performance on mobile devices, this approach proves effective for field applications. This research has practical implications for biodiversity conservation, herbal medicine, and sustainable resource management, offering a feasible solution for automated, real-time herbal plant identification. Future work will focus on expanding the dataset and refining the model to cover a broader range of species and improve robustness across varied ecological conditions.

Keywords: Herbal plant detection, machine learning, computer vision, Transformer Model, biodiversity conservation, species identification, real-time classification, sustainable resource management.