Investigation of Bispectral Co-Occurrence Matrices for Multispectral Data Analysis

Authors

  • V. A. Smetanin hysical-Technical Institute of the Ural Branch of RAS, Izhevsk
  • V. N. Milich hysical-Technical Institute of the Ural Branch of RAS, Izhevsk

Keywords:

multispectral images, neural network, image texture features, pattern recognition

Abstract

A system of image texture features developed on the basis of bispectral co-occurrence matrices is explored. A neural network is used for classifying decision over the features. Results of experiments with the real multispectral data are showed. The paper presents the established dependencies of quality for both the network learning and the recognition on the learning duration and the training extent.

Author Biographies

V. A. Smetanin, hysical-Technical Institute of the Ural Branch of RAS, Izhevsk

Post-graduate

V. N. Milich, hysical-Technical Institute of the Ural Branch of RAS, Izhevsk

PhD in Engineering

References

Сметанин В. А., Милич В. Н. Возможности использования текстурных признаков изображений для анализа материалов многозональной съемки // Материалы VII Всерос. науч.-техн. конф. «Приборостроение в XXI веке. Интеграция науки, образования и производства». – Ижевск : Изд-во ИжГТУ, 2012. – С. 484–489.

Haralick R. M., Shanmugam K., Distein It'sHak. Textural Features for Image Classification // IEEE Transactions on systems, man and cybernetics. – 1973. – Vol. SMC_3. – № 6. – P. 610−621.

Астафуров В. Г., Скороходов А. В. Классификация облаков по спутниковым снимкам на основе топологии нейронных сетей // Современные проблемы дистанционного зондирования Земли из космоса. – 2011. – Т. 8. – № 1. – С. 65–72.

Generalization of the Cooccurrence Matrix for Colour Images: Application to Colour Texture Classification / V. Arvis [et al.] // Image Analysis and Stereology. – 2004. – Vol. 23. – P. 63–72.

Хайкин С. Нейронные сети: полный курс : пер. с англ. – 2-е издание // М. : Вильямс, 2006. – 1104 с.

Published

15.12.2012

How to Cite

Smetanin В. А., & Milich В. Н. (2012). Investigation of Bispectral Co-Occurrence Matrices for Multispectral Data Analysis. Vestnik IzhGTU Imeni M.T. Kalashnikova, (4), 123–126. Retrieved from https://izdat.istu.ru/index.php/vestnik/article/view/2688

Issue

Section

Articles