Krylov V. Structural stochastic identification of geometrical objects and textures for automatizated control systems and information processing

Українська версія

Thesis for the degree of Doctor of Science (DSc)

State registration number

0502U000421

Applicant for

Specialization

  • 05.13.06 - Інформаційні технології

24-10-2002

Specialized Academic Board

Д 41.052.01

Odessa National Polytechnic University

Essay

The object is process of geometrical objects and textures noisy stability identification for automatizated control systems; the aims to design stochastic structural identification theoretical base for improving of processing efficiency and compression information rate; the methods are functional transforming theory, mathematical statistic, pattern recognition theory; the new approach lies in edge detection methods, which are the composition of edge improving and multichanel adaptive optimal detection, stochastic solutions and estimations domain rank processing methods, geometrical objects and textures signs calculation from contour inflection points (IF), noisy stability and local extremes indifferent fetcher fragment regular iteration search by Hilbert domain, quality , masking function, assignment estimation methods with involving of human visual system properties, the contour IF detection method by linear interpolation. The outcomes under the image recognition theory. This thesis has been applicated for automatizated nuclear assemblies TV control system (NPP), for ecological monitoring, for learning process. Sphere of use is image processing automatizated control system.

Files

Similar theses