Automatic Identification of Butterfly Species Based on Gray-Level Co-occurrence Matrix Features of Image Block

doi:10.1007/s12204-018-2013-y

Abstract

Abstract: In recent years, automatic identification of butterfly species arouses more and more attention in different areas. Because most of their larvae are pests, this research is not only meaningful for the popularization of science but also important to the agricultural production and the environment. Texture as a notable feature is widely used in digital image recognition technology; for describing the texture, an extremely effective method, graylevel co-occurrence matrix (GLCM), has been proposed and used in automatic identification systems. However, according to most of the existing works, GLCM is computed by the whole image, which likely misses some important features in local areas. To solve this problem, this paper presents a new method based on the GLCM features extruded from three image blocks, and a weight-based k-nearest neighbor (KNN) search algorithm used for classifier design. With this method, a butterfly classification system works on ten butterfly species which are hard to identify by shape features. The final identification accuracy is 98%.

Key words: automatic identification| butterfly species| gray-level co-occurrence matrix (GLCM)| features of image block

摘要： In recent years, automatic identification of butterfly species arouses more and more attention in different areas. Because most of their larvae are pests, this research is not only meaningful for the popularization of science but also important to the agricultural production and the environment. Texture as a notable feature is widely used in digital image recognition technology; for describing the texture, an extremely effective method, graylevel co-occurrence matrix (GLCM), has been proposed and used in automatic identification systems. However, according to most of the existing works, GLCM is computed by the whole image, which likely misses some important features in local areas. To solve this problem, this paper presents a new method based on the GLCM features extruded from three image blocks, and a weight-based k-nearest neighbor (KNN) search algorithm used for classifier design. With this method, a butterfly classification system works on ten butterfly species which are hard to identify by shape features. The final identification accuracy is 98%.

关键词: automatic identification| butterfly species| gray-level co-occurrence matrix (GLCM)| features of image block

CLC Number:

TP 391.4

XUE Ankang (薛安康), LI Fan* (李凡), XIONG Yin (熊吟). Automatic Identification of Butterfly Species Based on Gray-Level Co-occurrence Matrix Features of Image Block[J]. Journal of Shanghai Jiao Tong University (Science), 2019, 24(2): 220-225.

References 20

[1]	ZHOU Y. Monographia rhopalocerorum sinensium[M]. Zhengzhou, China: Henan Science and TechnologyPress, 1994 (in Chinese).
[2]	GASTON K J, O’NEILL M A. Automated speciesidentification: Why not? [J]. Philosophical Transactionsof The Royal Society B, Biological Sciences, 2004,359(1444): 655-667.
[3]	GASTON K J, MAY R M. Taxonomy of taxonomists[J]. Nature, 1992, 356: 281-282.
[4]	WEEKS P J D, O’NEILL M A, GASTON K J, et al.Species-identification of wasps using principal componentassociative memories [J]. Image and Vison Computing,1999, 17(12): 861-966.
[5]	WEEKS P J D, O’NEILL M A, GASTON K J, et al.Automating insect identification: Exploring the limitationsof a prototype system [J]. Journal of AppliedEntomology, 2010, 123(1): 1-8.
[6]	HOPKINS G W, FRECKLETON R P. Declines in thenumbers of amateur and professional taxonomists: Implicationsfor conservation [J]. Animal Conservation,2010, 5(3): 245-249.
[7]	ARBUCKLE B, SCHR¨OEDER S, STEINHAGE V,et al. Biodiversity informatics in action: Identificationand monitoring of bee species using ABIS[C]//Proceedings of 15th International Symposium Informaticsfor Environment Protection. Marburg, Germany:Metropolis Verlag, 2001: 425-430.
[8]	YAO Q, LV J, LIU Q J, et al. An insect imaging systemto automate rice light-trap pest identification [J].Journal of Integrative Agriculture, 2012, 11(6): 978-985.
[9]	KAYA Y, KAYCI L. Application of artificial neuralnetwork for automatic detection of butterfly speciesusing color and texture features [J]. Visual Computer,2014, 30(1): 71-79.
[10]	KANG S H, CHO J H, LEE S H. Identification of butterflybased on their shapes when viewed from differentangles using an artificial neural network [J]. Journal ofAsia-Pacific Entomology, 2014, 17(2): 143-149.
[11]	WANG J N, JI L Q, LIANG A P, et al. The identificationof butterfly families using content based image retrieval[J]. Biosystems Engineering, 2012, 111(1): 24-32.
[12]	KAYA Y, KAYCI L, TEKIN R. A computer vision systemfor the automatic identification of butterfly speciesvia Gabor-Filter-Based texture features and extremelearning machine: GF+ELM [J]. TEM Journal, 2013,2(1): 13-20.
[13]	VANITHA R, PREMANANDA R. Content based imageretrieval using color and texture feature [J]. InternationalJournal of Advanced Technology in Engineeringand Science, 2014, 2(6): 308-318.
[14]	HUANG K Y. Application of artificial neural networkfor detecting Phalaenopsis seedling diseases using colorand texture features [J]. Computers and Electronics inAgriculture, 2007, 57(1): 3-11.
[15]	KANG S H, JEON W, LEE S H. Butterfly speciesidentification by branch length similarity entropy [J].Journal of Asia-Pacific Entomology, 2012, 15(3): 437-441.
[16]	LI H F, CHAI Y, LI Z F. A new fusion scheme for multifocusimages based on focused pixels detection [J].Machine Vision and Applications, 2013, 24(6): 1167-1181.
[17]	LI H F, LIU X K, YU Z T, et al. Performance improvementscheme of multifocus image fusion derivedby difference images [J]. Signal Processing, 2016, 128:474-493.
[18]	LI H F, LI X S, YU Z T, et al. Multifocus image fusionby combining with mixed-order structure tensorsand multiscale neighborhood [J]. Information Sciences,2016, 349: 25-49.
[19]	LI H F, YU Z T, MAO C L. Fractional differentialand variational method for image fusion and superresolution[J]. Neurocomputing, 2016, 171: 138-148.
[20]	LI H F, QIU H M, YU Z T, et al. Multifocus imagefusion via fixed window technique of multiscale imagesand non-local means filtering [J]. Signal Processing,2017, 138: 71-85.