Classification of Herbal Plant and Comparative Analysis of SVM and KNN Classifier Models on the Leaf Features Using Machine Learning

Abstract

Herbal plant species classification is the most widely used area in the field of research in image processing, machine learning and deep learning. To identify and classify plant, leaf feature is used, as it is available throughout the year. Different sizes and shapes of leaf from single plant are taken for classification purpose. In this paper, an automated system is discussed in which various features of leaf shape are calculated such as length, width, area of leaf, leaf perimeter in pixels, area of rectangle enclosing leaf, leaf percentage in rectangle. The image of leaf has been taken with the mobile camera and then resized it for fast evaluation and calculating the features. Preprocessing technique is applied to convert colored images to gray scale and binary image for extracting their shape of leaf easily. In this paper, two classifiers—SVM model (support vector machine) and KNN model (k-nearest neighbor) are used to classify the plants to obtain the optimum accuracy and the results are discussed.