Automatic classification of plants based on their leaves

• Published in: Biosystems engineering Vol. 139; pp. 66 - 75
• Main Authors: Aakif, Aimen, Khan, Muhammad Faisal
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.11.2015
• Subjects: Algorithms; Artificial neural networks; Classification; Classification of plants; Feature extraction; Fourier analysis; Leaf segmentation; Learning theory; Leaves; Neural networks; Plants (organisms); Shape defining feature; Feature extraction; Classification of plants; Shape defining feature; Leaf segmentation; Artificial neural networks
• ISSN: 1537-5110; 1537-5129
• DOI: 10.1016/j.biosystemseng.2015.08.003

The proposed algorithm identifies a plant in three distinct stages i) pre-processing ii) feature extraction iii) classification. Different leaf features, such as morphological features, Fourier descriptors and a newly proposed shape-defining feature, are extracted. These features become the input vector of the artificial neural network (ANN). The algorithm is trained with 817 samples of leaves from 14 different fruit trees and gives more than 96% accuracy. To verify the effectiveness of the algorithm, it has also been tested on Flavia and ICL datasets and it gives 96% accuracy on both the datasets. •We propose an algorithm to classify a plant on the basis of its features.•In addition to traditional features, we propose a new ‘shape defining feature’.•Artificial neural network with back propagation has been used as a classifier.•Different algorithms for analysis have been compared.