Medicinal Plants Attribute Detection by Deep learning Image Processing Techniques

The identification of medicinal plant species leads to many benefits to wide range of stakeholders such as, forestry services, taxonomists, doctors, and pharmaceutical labs, organizations that protect endangered species, the government, and the general public. Important plant species are disappearing around the world on a daily basis. Steps include plant identification, restoration, and protection in order to control the phenomenon. Typically, leaves are used to identify plants. Convolutional Neural Network which is the widely used architecture for image base classification motivated us to creating automated methods for classifying various plant species and identify medicinal attributes of the plants. In this paper, using computer vision and machine learning approaches, a fully automated front end based software system has been designed and described for achieving this. Each leaf's length, width, perimeter, area, number of vertices, color, perimeter, and hull area were all just a few of the many features that were taken out. Then, from these properties, a number of derived features were calculated. The native people are expected to get more knowledge about medicinal leaves with the use of this work for the automatic recognition of medicinal leaves. Additionally, it helps taxonomists in the development of more effective methods for identifying species, which will greatly help in the preservation of endangered species. The utilization of various plant traits ranges from the most basic, such as leaf color and form, to the most complicated, such as tissue and cell structure are applied. The results obtained with these analytics gives 98% accuracy in 100 epochs.