ALGORITHM OF FORMING AND SELECTING OF INFORMATIVE FEATURES IN CORRELATION EXTREME NAVIGATION SYSTEM DATABASE

Authors

  • M. P. Mukhina National Aviation University
  • A. P. Prymak National Aviation University
  • A. N. Babeniuk National Aviation University

DOI:

https://doi.org/10.18372/1990-5548.47.10292

Keywords:

Feature selection, maximum entropy method, correlation weighting, Speed-Up Robust Feature, template informativity

Abstract

Hierarchical approach to data fusion in complex navigation system is proposed. Low level ofdata processing in correlation extreme navigation system includes forming and selecting the most informativerelevant features. Advanced algorithm is therefore developed based on weighting correlationsampling of template features. Practical efficiency of algorithm is researched based on Speed-Up RobustFeature method of feature points detection and description. Minimization of feature points number intemplate without decreasing the quality of matching is observed along with reduction of computationalcosts

Author Biographies

M. P. Mukhina, National Aviation University

Ph.D. Associate professor. Educational-Scientific Institute of Information Diagnostic Systems

A. P. Prymak, National Aviation University

6th year of study student

A. N. Babeniuk, National Aviation University

6th year of study student

References

M.P. Mukhina. “Prospects of development correlation-extreme modern navigation systems”. AVIA-2013: Materials of the XI International scientific technical conference. Kyiv: NAU, 2013. pp. 22.1–22.4.

V. К. Baklitskiy. Correlation-extreme methods of navigation and targeting. Tver.: TO "Book Club", 2009, 360 p

Eyink G. L., Kim S. A maximum entropy method for particle filtering //Journal of statistical physics. – 2006. – vol. 123. no. 5. pp. 1071–1128.

Herbert Bay, Tinne Tuytelaars, Luc Van Gool. SURF: Speeded Up Robust Features. – Katholieke Universiteit Leuven. 2007. 13 p

Koller D., Sahami M. Toward optimal feature selection. – Machine Learning: Proceedings of the 13-th International Conference on Machine Learning. 1996. Code of SURF listing in MATLAB. Available from Internet:

http://www.mathworks.com/matlabcentral/fileexchan ge/28300-opensurf-including-image-warp

Downloads

Issue

Section

TRANSPORT SYSTEMS