Commonly, classic plant identification methods use dichotomous or multi-access keys that compare characteristics of the leaves, asking if they are lobed, unlobed, simple or compound, among others leaf features. However, in the literature very little attention has been paid to make an automatic distinction of leaves using such features. In this paper, we contribute to fill this gap. We propose a novel method to differentiate between types of leaves. The proposal is invariant to rotation and also to scaling. In order to show the effectiveness of the proposal, we tested it with more than 1,900 images of leaves which are publicly available on the Internet, achieving correct identification rates greater than 86%.