Objects (signals) identification using estimated information entropy of two-dimensional monochrome images
- Details
- Category: Information technologies, systems analysis and administration
- Last Updated on 30 October 2015
- Published on 30 October 2015
- Hits: 3618
Authors:
S.I. Melnychuk, Cand. Sci. (Tech.), Associate Professor, Ivano-Frankivsk National Technical University of Oil and Gas, Professor of the Department of Computer Technology in Systems of Controlled and Automation, Ivano-Frankivsk, Ukraine
Abstract:
Purpose. Improving the efficiency of the objects identification computational methods using their monochrome images through the development of efficient technologies, methods and means of processing of digital representations (signals) with variable and changing information entropy. Development of the object identification method based on processing of probabilistic representations and projections that are formed using statistical evaluation of information entropy values of related fragments of objects characteristics. Investigation of the efficiency of correlation based processing of signal entropy representations and projections that are implemented based on the proposed approach.
Methodology. The main scientific results and conclusions were made on a basis of information theory, theory of signals and systems, probability theory and mathematical statistics, methods of spectral and statistical analysis, digital signal processing and based on results of computational modeling.
Findings. Based on computational modeling it was discovered that the proposed computational method allows achieving of unambiguous identification under information loss caused by image distortion that does not exceed 87%. Besides, the independence of the proposed approach from inverse representation in the case of monochrome images was confirmed.
Originality. The improved computational projection method of the object identification for which construction of projections is proposed to be done by statistical evaluation of information entropy values of corresponding fragments of 2D binary matrices of objects monochrome representations.
Practical value. The proposed method allows increasing the correctness of identification in comparison with known projection based computational methods.
References:
1. Intelligent Security Systems, “Container Recognition ‒ SecurOS Cargo”, (2014), Available at: http://cams.isscctv.com/products/intelligent/1716/, (Access date April 22, 2014)
2. OV-SOFT, “Face numbers of rail cars”, Available at: http://www.ovsoft.ru/products/davis/davis-railway.html, (Access date April 22, 2014)
ОВ-СОФТ. Система розпізнавання номерів залізничних вагонів [Електронний ресурс]. – Режим доступу: http://www.ovsoft.ru/products/davis/davis-railway.html
3. Potapov, A.A., Guljaev, Ju.V., Nikitov, S.A., Pakhomov, A.A. and German, V.A. (2008), Noveyshye metody obrabotki izobrazheniy [The Latest Image Processing Techniques], FIZMATLIT, Moscow, Russia.
Новейшие методы обработки изображений / [Потапов А.А., Гуляев Ю.В., Никитов С.А. и др.] – М.: ФИЗ-МАТЛИТ, 2008. – 496с.
4. Sablina, V.A., Novikov, A.I. Nikitfrov, M.B. and Loginov, A.A. (2013), “An Approach to the Superimposition Problem in Multispectral Computer Vision Systems”, Proc. of the 2nd Mediterranean Conference on Embedded Computing (MECO), Budva, Montenegro, IEEE Catalog Number CFP1339T-ART, pp. 117–120.
5. Kong, T.Y. and Rosenfeld, A. (1989), “Digital topology: Introduction and Survey”, Comp. vision, graphics and image processing, Vol. 48. pp. 357–393.
6. Melnychuk, S.I. (2012), Patent 103281 Ukraine, MPK (2006) G06K 9/64. Identifying objects method by probabilistic characteristics of images fragments, request No. a 2012 12161; applied on October 23, 2012, published on September 25, 2013, Bulletin no. 18.
Патент 103281 Україна, МПК(2006) G06K 9/64. Спо-сіб ідентифікації об’єктів за ймовірнісними характери-стиками фрагментів їх зображень / Мельничук С.І. (Україна). – заявка № а 2012 12161; заявл. 23.10.2012; опубл. 25.09.2013, Бюл. № 18.
7. Melnychuk, S.I. (2013), “Identifying objects by probabilistic characteristics of fragments of monochrome images”, Proc. of the 11th International Scientific and Practical Conference: Mathematical and program support of intelligent systems, Dnipropetrovsk, November 20-22, 2013, TOV Royal Print, Dnipropetrovsk pp. 159–160.
Мельничук С.І. Ідентифікація об’єктів за ймовірнісними характеристиками фрагментів їх монохромних зображень: XI міжнародна науково-практична конференція: Математичне та програмне забезпечення інтелектуальних систем, м. Дніпропетровськ, 20–22 листопада 2013 р. / С.І Мельничук – Дніпропетровськ: ТОВ „Роял Принт“, 2013. – С. 159–160.
8. MathCAD12. Available at: http://mathcad.name/Glava_01/Index03.htm, (Access date April 22, 2014)
MathCAD12. [Електронний ресурс] – Режим доступу: http://mathcad.name/Glava_01/Index03.htm.
2015_03_melnychuk | |
2015-10-29 714.07 KB 933 |