classification of plant leaves using morphological features and Hu moments

Abstract

Plants are an important in our ecosystems. Their identification and classification has always been a matter of interest for the botanists as well as for humans. With huge number of plant species, only a tiny part of the plants is known. The leaves of the plant are so important because they carry a lot of information about the plant species. The aim of this work is to automatically classify leaves into different classes through their images. The advancement in image processing and machine learning has made this a quick and easy process. Specifically, we describe leaf images using morphological features together with Hu moments. Then, we use the KNN and Naive Bayes classifiers to first learn the visual characteristics of different leaf classes and then to decide for test leaves the classes they belong to. In order to truly investigate the performance of our system, we carried out many experiments on the well-known Flavia dataset. Results reported at the end of this work show promising results.