期刊
JOURNAL OF KING SAUD UNIVERSITY-COMPUTER AND INFORMATION SCIENCES
卷 34, 期 1, 页码 1361-1369出版社
ELSEVIER
DOI: 10.1016/j.jksuci.2018.09.018
关键词
Leaf venation; Plant identification; Multimodal plant; Leaf spectral signature
资金
- UGC-MANF fellowship
Automatic classification of plants using image processing and machine learning techniques based on leaf venation patterns and spectral signatures can achieve high accuracy, with feature fusion technique outperforming non-imaging spectral signatures features. This approach shows promise for efficient plant identification.
Automatic classification of the plants is growing area of association with computer science and Botany, it has attracted many researchers to subsidize plant classification using image processing and machine learning techniques. Plants can be classified using number of traits such as leaf color, flowers, leaves, roots, leaf shape, leaf size etc. highly depends upon feature selection methods. However extraction of features from selected trait is most significant state in classification. State-of-the-art classification can be achieved by using leaf characteristics such as leaf venation patterns, leaf spectral signatures, leaf color, leaf shape, etc. This paper describes multimodal plant classification system using leaf venation patterns and its spectral signatures as a significant features. This paper shows that the feature fusion can be used to achieve efficient plant identification. The accuracy of identification for leaf spectral data, leaf venation features and HOG features is validated, it signifies that feature fusion technique performs better than that of non-imaging spectral signatures features only, with recognition result of 98.03% GAR and 93.51% GAR, respectively. (c) 2018 The Authors. Production and hosting by Elsevier B.V. on behalf of King Saud University. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据