PLANTS SPECIES IDENTIFICATION USING COMPUTER VISION TECHNIQUES

Abstract

Plants are quite important component in our ecosystem. Botanists need to identify plants type for different targets, for example distinguishing the ones which can be used for medical purposes. Traditionally, botanists identify plants manually by using cellular and biological characteristics, which is, in fact, a tedious and time consuming process. Therefore, designing an automatic system, which is capable to identify the different types of plants, is highly recommended. In this paper, we propose a fully automatic method for leaves classification based on computer vision techniques. Instead of extracting the cellular characteristics of plants, our proposed method recognize the type of the plant from the visual features i.e., characteristics which is extracted from a leaf image. The used features include the leaf length, width and diameter. The proposed method is fully automatic, as it doesn’t require any human intervention. In addition, it allows persons who are not familiar with the biology domain to recognize the plants type. To prove the efficiency of the proposed system, we conduct experiments on Flavia dataset which assembles 1907 leaf images of 32 types of plants. Experimental results show promising results and an accuracy of 94% has been reached.