Introducing new statistical shape based and texture feature extraction methods in the plant species recognition system
Plant species recognition system (PSRS) plays important roles in automated agricultural systems. Usually shape and texture based techniques are used at the same time in object recognition problems. Shape based features are important features in image processing literature and also as a method that t...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Federation of Engineering Institutions of Islamic Countries
2013
|
| Online Access: | http://psasir.upm.edu.my/id/eprint/18192/1/Introducing%20new%20statistical%20shape%20based%20and%20texture%20feature%20extraction%20methods%20in%20the%20plant%20species%20recognition%20system.pdf |
| _version_ | 1825946148164075520 |
|---|---|
| author | Seyed Mohammad Hussein, Ahmad, Siti Anom Hassan, Mohd Khair Ishak, Asnor Juraiza |
| author_facet | Seyed Mohammad Hussein, Ahmad, Siti Anom Hassan, Mohd Khair Ishak, Asnor Juraiza |
| author_sort | Seyed Mohammad Hussein, |
| collection | UPM |
| description | Plant species recognition system (PSRS) plays important roles in automated agricultural systems. Usually shape and texture based techniques are used at the same time in object recognition problems. Shape based features are important features in image processing literature and also as a method that the human visual system applies to recognize objects Therefore that is a feature representation considered here. In this paper, features such as median, mean, variance and standard deviation (SD) is engaged in shape representation. In addition to shape based features, ROI (region of interest) -entropy average (REA) is introduced to extract texture base features. These two methods are tested on the leaf samples to evaluate the performance of two proposed methods. In addition, in this research NNUGA used for PSRS problem which could increase the accuracy of feed forward neural networks. The results show the outperformance of the two proposed methods for image processing and optimized classifier for classification part. As the classification result, radial basis neural networks (RBFNN), feed forward neural networks (FFNN), neural networks using genetic algorithm (NNUGA) shows 100%, 93%, 97.3% of accuracy respectively .
|
| first_indexed | 2024-03-06T07:42:32Z |
| format | Article |
| id | upm.eprints-18192 |
| institution | Universiti Putra Malaysia |
| language | English |
| last_indexed | 2024-03-06T07:42:32Z |
| publishDate | 2013 |
| publisher | Federation of Engineering Institutions of Islamic Countries |
| record_format | dspace |
| spelling | upm.eprints-181922016-01-28T04:56:30Z http://psasir.upm.edu.my/id/eprint/18192/ Introducing new statistical shape based and texture feature extraction methods in the plant species recognition system Seyed Mohammad Hussein, Ahmad, Siti Anom Hassan, Mohd Khair Ishak, Asnor Juraiza Plant species recognition system (PSRS) plays important roles in automated agricultural systems. Usually shape and texture based techniques are used at the same time in object recognition problems. Shape based features are important features in image processing literature and also as a method that the human visual system applies to recognize objects Therefore that is a feature representation considered here. In this paper, features such as median, mean, variance and standard deviation (SD) is engaged in shape representation. In addition to shape based features, ROI (region of interest) -entropy average (REA) is introduced to extract texture base features. These two methods are tested on the leaf samples to evaluate the performance of two proposed methods. In addition, in this research NNUGA used for PSRS problem which could increase the accuracy of feed forward neural networks. The results show the outperformance of the two proposed methods for image processing and optimized classifier for classification part. As the classification result, radial basis neural networks (RBFNN), feed forward neural networks (FFNN), neural networks using genetic algorithm (NNUGA) shows 100%, 93%, 97.3% of accuracy respectively . Federation of Engineering Institutions of Islamic Countries 2013-08-04 Article PeerReviewed application/pdf en http://psasir.upm.edu.my/id/eprint/18192/1/Introducing%20new%20statistical%20shape%20based%20and%20texture%20feature%20extraction%20methods%20in%20the%20plant%20species%20recognition%20system.pdf Seyed Mohammad Hussein, and Ahmad, Siti Anom and Hassan, Mohd Khair and Ishak, Asnor Juraiza (2013) Introducing new statistical shape based and texture feature extraction methods in the plant species recognition system. International Journal of Engineering and Technology, 10 (2). pp. 56-61. ISSN 1823-1039 http://www.ijet.feiic.org/index.php/volume-10-2013 |
| spellingShingle | Seyed Mohammad Hussein, Ahmad, Siti Anom Hassan, Mohd Khair Ishak, Asnor Juraiza Introducing new statistical shape based and texture feature extraction methods in the plant species recognition system |
| title | Introducing new statistical shape based and texture feature extraction methods in the plant species recognition system |
| title_full | Introducing new statistical shape based and texture feature extraction methods in the plant species recognition system |
| title_fullStr | Introducing new statistical shape based and texture feature extraction methods in the plant species recognition system |
| title_full_unstemmed | Introducing new statistical shape based and texture feature extraction methods in the plant species recognition system |
| title_short | Introducing new statistical shape based and texture feature extraction methods in the plant species recognition system |
| title_sort | introducing new statistical shape based and texture feature extraction methods in the plant species recognition system |
| url | http://psasir.upm.edu.my/id/eprint/18192/1/Introducing%20new%20statistical%20shape%20based%20and%20texture%20feature%20extraction%20methods%20in%20the%20plant%20species%20recognition%20system.pdf |
| work_keys_str_mv | AT seyedmohammadhussein introducingnewstatisticalshapebasedandtexturefeatureextractionmethodsintheplantspeciesrecognitionsystem AT ahmadsitianom introducingnewstatisticalshapebasedandtexturefeatureextractionmethodsintheplantspeciesrecognitionsystem AT hassanmohdkhair introducingnewstatisticalshapebasedandtexturefeatureextractionmethodsintheplantspeciesrecognitionsystem AT ishakasnorjuraiza introducingnewstatisticalshapebasedandtexturefeatureextractionmethodsintheplantspeciesrecognitionsystem |