Plant Species Classification Using Leaf Shape and Texture

It is of vital importance as well as a great challenge to recognize plant species on the earth planet, from which human beings can benefit much. Thus it would be useful to design a convenient and effective image classification method to automatically classify different species. To reach this goal, in this paper we propose a new method to generate the feature space that combines local texture features using wavelet decomposition and co-occurrence matrix statistics and global shape features to describe the collected plant leaves. Finally, experiments are conducted using SVM (Support Vector Machine) classifiers to classify the different species. Experimental results show that our proposed methods achieve accuracy over 93.8% using a data set with over 1900 leaves from 32 species, exceeding most approaches that have been proposed.