Plant leaf species identification using Curvelet transform

Abstract

In this paper, a novel approach for feature extraction from natural image such as plant leaf is proposed for automated living plant species recognition useful for botanical students in their research for plant species identification. A new multi-resolution and multidirectional Curvelet transform is applied on subdivided leaf images to extract leaf information, mathematically so that the orientation of the object in the image does not matter and which also increase the accuracy rate. These coefficients will be the input to a trained SVM classifier to classify the result. Compared to other exiting methods and tools in this field of plant species recognition the proposed system gives a higher accuracy rate of around 95.6% with 624 leaf dataset.