Avtomatsko prepoznavanje prikaza segmentnega prikazovalnika s pomočjo računalniškega vida

MUBI, MIHA (author), Agrež, Dušan (mentor)

Računalniški vid je v zadnjem času pomembno orodje na področju avtomatizacije posameznih postopkov. To še posebej velja v primerih, ko zaradi dolgotrajnosti postopka človeška zbranost pada. Tudi v podjetju Iskraemeco d.d., v oddelku Raziskave in razvoj smo prišli do spoznanja, da bi nam tako orodje prišlo še kako prav pri testiranju vgrajene programske kode, ki teče na elektronskih števcih električne energije. Za zajem in ustrezno obdelavo slik smo pri razvoju projekta uporabili preverjene algoritme. To so algoritmi iz odprtokodne knjižnice OpenCV, katere smo v celoto povezali s pomočjo programskega jezika Python. Vseeno pa je bilo potrebno nekaj programskih funkcij razviti tudi samostojno. Za zanesljivo zajemanje slik v realnem času smo uporabili kamero Logitech c920. Na koncu smo na novo razvito orodje tudi testirali. To smo izvedli s pomočjo avtomatskih testov, ki smo jih razvili s pomočjo programskega okolja Visual Studio 2017 in programskega jezika C#. Tu smo že v fazi testiranja odkrili prve hrošče na vgrajeni programski opremi števca in tako prikazali uporabnost orodj.

• Language: Slovenian
• Keywords: računalniški vid, elektronski števec električne energije, knjižnica OpenCV, Python, spletna kamera Logitech c920, Visual Studio 2017, C#
• Work type: Master's thesis/paper
• Organization: FE - Faculty of Electrical Engineering
• Year: 2018
• PID: 20.500.12556/RUL-103811
• Publication date in RUL: 26.09.2018

Secondary language

• Language: English
• Title: Automatic recognition of the screen segments of a display using computer vision

Abstract

Computer vision is an important tool in automatization of individual procedures. This is especially true in cases where the human concentration falls due to the length of the procedure. In the Iskraemeco d.d. company, in the Research and Development section, we came to the realization that the tool would be just as useful for testing the firmware that runs in electronic meters of electricity. In this project, we used verified algorithms to capture and properly process images. This is OpenCV open-source library, which we have integrated into the whole using the Python programming language. Nevertheless, some program’s functions had to be developed by ourselves. We used the Logitech c920 webcam for real time images capture. In the end, we also tested the newly developed tool. This was done using automated tests that we developed using the Visual Studio 2017 programming environment and the C# programming language. In this phase, we have already discovered the first bugs of meter’s firmware and so showing the utility of the tool.

• Keywords: computer vision, electronic meter of electricity, OpenCV library, Python, Logitech c920 webcam, Visual Studio 2017, C#