Odborná spolupráce společných přeshraničních vědecko-výzkumných skupin

Spolupracující instituce se dohodly na zřízení dvou výzkumných skupin

1. Téma: Metody číslicového zpracování obrazu

Garant TU Ostrava – ing. Richard Velička, Ph.D.

Členové TU Ostrava

Pedagogové: doc. P. Bilík, Ph.D., ing. R. Velička, Ph.D.

Doktorandi a studenti: Ing. Martin Kryl, Ing. Lukáš Šoustek

Členové PO Opole

Pedagogové: prof. M. Tomaszewski, ing. P. Michalski, Ph.D., ing. B. Ruszczak, Ph.D., ing. M. Kamiński, Ph.D., ing. R. Gasz, Ph.D.

Doktorandi a studenti: ing. J. Osuchowski, M.Sc., ing. A. Różańska, M.Sc.

Aktivity:

• Společná setkání výzkumných skupin, střídavě v CZ a PL, kde budou diskutovány odborné aktivity doktorandů obou stran
• Specializovaná laboratorní práce doktorandů

Odborná náplň je uvedena v následujícím textu. Text je uveden v angličtině, protože komunikačním jazykem je převážně angličtina.

• Our goal is to make universal solution and rapid-development module for machine vision industrial tasks which refer to searching for imperfections and faulty parts of products. Since we needed to start on something particular, so we could aply and test alogrithms, we chose wood industry which is “rich“ on non-ideal structures and imperfections like knots, scratches, craks etc. Moreover surface quality and wood type classification tasks could be done on wood as well.
• As first, we developed wood classification application which is based on machine learning algorithms Support Vector Machine (SVM) and K-nearest Neighbor (KNN). Interface is in this case designed in the matter, whole testing process including dataset creation, algorithm computation and testing, can be completed within short period of time.

• Second approach aim to local error detection. In first phase just knots recognision have been aplied. Other quality problems will be included afterwards. Convolutional Neural Networks were in this case chosen as the most universal solution.

• Our goal is to development method of virtual photographs as a tool to support the work of technicians in the process of analyzing the photographic material collected during annual areal inspections of the lines. The primary objective of the virtual photograph method is to limit the number of analyzed photographs by qualified personnel to only those photographs which indicate the potential risk of inconsistencies with the model - lack of an element or deformation. This method incorporates mechanisms drawn from screening analysis.
• In order to achieve a correctly working method, the study describes the adopted principles of photograph evaluation, incorporates a series of analyses regarding the correct determination of irregularities in the latticework structure.

• Due to enormuse amount of digital data generated during a high voltage line insulator inspection, there is a high need to autmate this process. Our goal is to construct a system which can analyzed photos from UAV vehicles, to support insulators diagnosis process. The first step in this process is to detect insulators in digital images. To achieve this goal the proposed system utilizes a Convolutional Neural Networks (CNN) to predict bounding boxes which contain only insulators.

• Used computer vision and image processing to distinguish various human emotions and to categorize them. Categorized emotions are used for future AI as input values for it to be able to correctly recognize an emotion and for AI to be able to reply with emotions of it own.

2. Téma: Analiza systemów technicznych z wykorzystaniem metod przetwarzania obrazu, narzędzi symulacyjnych oraz systemów pomiarowych

Garant PO Opole - ing. Rafał Gasz, Ph.D.

Členové TU Ostrava

Pedagogové: doc. P. Bilík, Ph.D., ing. R. Velička, Ph.D., prof. P. Santarius, CSc.

Doktorandi a studenti:

Členové PO Opole

Pedagogové: prof. M. Tomaszewski, ing. P. Michalski, Ph.D., ing. R. Gasz, Ph.D., ing. K. Wróbel, Ph.D., ing. K. Górecki, Ph.D., ing. Ł. Dzierżanowski, Ph.D., ing. M. Lasar, M.Sc.

Doktorandi a studenti: ing. S. Szymocha, M.Sc., ing. R. Gabor, M.Sc., ing. K. Kozioł, M.Sc.

Odborná náplň

• Analysis of the influence of constructional and material parameters on the operation of a passive magnetic gear. In the work author use numerical modeling methods to model, examine and evaluate the magnetic field and vibrations of the magnetic gear. The work includes: Analysis of magnetic forces acting on the modulator poles, Determination of the impact of material parameters on the stiffness of the gear construction, Design and execution of gears with improved parameters.
• Measurement of air quality using UAV. Measurement of the UAV unit enables quick collection of data on a large area. The challenge in these tests is the selection of sensors and an algorithm optimizing the location of measuring points.

Struktura výzkumných skupin a odborná náplň byla diskutována na semináři dne 24. dubna 2019 v Hotelu Hluboký dvůr, Hlubočky.