An Automated System for Medicinal Plant Identification and Usage Recommendation

Abstract

Automatic plant image recognition is far and away the most likely answer to the botanical taxonomic gap, which has drawn more attention from the computing world as well as the botany community. More complex and sophisticated models for automatic plant picture identification have emerged in tandem with the continuous development of machine learning technology. Since medicinal plants are attracting increasing attention from the pharmaceutical sector due to their reduced side effect profile and significantly lower price compared to modern-day pharmaceuticals, many researchers have shown great interest in their regular recognition. Future research and development can go in a number of directions to create a reliable classifier that can recognise therapeutic plants in real time. After been recently applied to leaf imaging, this study examined a number of successful Deep Learning (DL) techniques for plant classification. This research study considered some DL classifiers referred to as Convolutional Neural Network (CNN) algorithm, and explained the image pre-processing techniques utilized to identify leaf and obtain meaningful leaf features. The classic plant properties (vein, shape, texture, and a variety of other characteristics) are used by these DL classifiers to operate or categorise leaf images. Additionally, provide the herb's scientific name, description, and usage. And then, with an increased accuracy rate fetch the results regarding herb usage. The system that was developed had a better accuracy rate, according to the test results.