Algoritem strojnega vida za zaznavanje objektov v delovnem področju robota

Kovač, Jan (author), Šimic, Marko (mentor), Herakovič, Niko (comentor)

Pri delu z roboti je večina nesreč posledica napačnega vedenja operaterja, ki vstopi v delovno področje robota. Da zmanjšamo vpliv človeškega faktorja smo v zaključni nalogi iskali možne koncepte strojnega vida, ki s pomočjo RGB kamere, ki je montirana na stropu, nadzirajo delovno področje robota in ob vstopu operaterja v delovno področje sprožijo varnostne funkcije. S pomočjo validacije posameznih konceptov smo izbrali najbolj ustreznega, ki je zanesljivo zaznal nevarnosti in zagotovil varnost operaterju. Koncept je enostavno implementirati v obstoječe sisteme.

• Language: Slovenian
• Keywords: strojni vid, zaznavanje, konvolucijske nevronske mreže, industrijski robot, python, varnost
• Work type: Final paper
• Typology: 2.11 - Undergraduate Thesis
• Organization: FS - Faculty of Mechanical Engineering
• Place of publishing: Ljubljana
• Publisher: [J. Kovač]
• Year: 2020
• Number of pages: XIV, 33 f.
• PID: 20.500.12556/RUL-117732
• UDC: 681.5:004.9:007.52(043.2)
• COBISS.SI-ID: 23956227
• Publication date in RUL: 23.07.2020

KOVAČ, Jan, 2020, Algoritem strojnega vida za zaznavanje objektov v delovnem področju robota [online]. Bachelor’s thesis. Ljubljana : J. Kovač. [Accessed 25 March 2025]. Retrieved from: https://repozitorij.uni-lj.si/IzpisGradiva.php?lang=eng&id=117732

Secondary language

• Language: English
• Title: Machine vision algorithm for object detection in robot's workspace

Abstract

When working with robots, most accidents are due to the misbehavior of the operator entering the robot’s work area. In order to reduce the influence of the human factor, we researched/examined possible concepts of machine vision, which, with the help of an RGB camera mounted on the ceiling, monitors the robot's work area and triggers security functions when the operator enters the work area. Through the validation of individual concepts, we selected the most appropriate one, which reliably detected hazards and ensured safety for the operator. The concept is easy to implement into existing systems.

• Keywords: machine vision, detection, convolutional neural networks, industrial robot, python, security