Medicinal Leaves Classification Using Random Forest and AdaBoost

The identification of medicinal plants is crucial for both biodiversity preservation and the creation of potent medications. Machine learning techniques will be employed to effectively classify and differentiate medicinal plant leaves based on their unique characteristics, enabling accurate species identification. The leaf images from six different species namely Neem, Tulasi, Peepal, Mint, Basale, Fenugreek have been collected. A comprehensive set of features, including shape features and texture features will be computed from the preprocessed images. In this project, the application of machine learning algorithms such as Random Forest and AdaBoost, for the classification of medicinal plant leaves will be being analyzed. The Random Forest classifies the images with and accuracy of 93.15% and the AdaBoost with an accuracy of 84.93%. Through the integration of some algorithms, this study seeks to provide a comprehensive framework for effectively categorizing different medicinal plant species based on their distinct features.

I. Introduction

Plants are integral to our lives, providing essential resources such as food, medicine, shelter, and oxygen. Among them, medicinal plants hold immense value for their therapeutic properties, which have been harnessed for centuries by various cultures around the world. However, accurately recognizing and identifying medicinal plants can be challenging due to their diverse morphological characteristics, geographical variations, and complex chemical compositions. Leaves play a crucial role in the classification of plants due to their unique and diverse characteristics. They serve as key identifiers in botanical studies, providing important information about the plant's species, health, and environmental adaptations. The shape, size, color, texture, venation pattern, and arrangement of leaves can vary significantly between different plant species, making them valuable features for distinguishing between plants. Additionally, leaves are often the most accessible and abundant plant parts, making them ideal for study and identification purposes. Their importance in plant classification extends beyond visual cues; leaves also contribute to the overall health and functioning of plants through processes such as photosynthesis and transpiration. As such, understanding the characteristics and variations of leaves is essential for accurately identifying and categorizing plant species, aiding in ecological studies, biodiversity conservation, and agricultural practices.