Zaznavanje barve in orientacije izdelkov s strojnim vidom v robotizirani stregi in montaži

diplomsko delo Visokošolskega strokovnega študijskega programa I. stopnje Strojništvo

Aljaž Mugerli (Author), Marko Šimic (Mentor), Niko Herakovič (Co-mentor)

Abstract

Diplomska naloga predstavlja razvoj sistema za zaznavo barve in orientacije kocke z uporabo računalniškega vida v kombinaciji z robotskim sistemom za strego sestavnih delov. Takšne sisteme imenujemo sistem strojnega vida, omogočajo avtomatizirano zaznavanje elementov in učinkovito ter natančno strego ali montažo. V prvem delu naloge so predstavljeni strojni vid, zaznava in obdelava barvne slike ter robotika. Pri strojnem vidu je opisan koncept delovanja strojnega vida in najbolj uporabljene metode zaznavanja. Nato je predstavljeno delovanje senzorja, ki je uporabljen v sistemu strojnega vida in ključni elementi za zaznavo slike, kot so leča, zaslonka in osvetlitev, bolje pa je predstavljen sistem za zaznavo barve na digitalni sliki. V drugem delu sta opisani strojna in programska oprema proizvajalca Dobot, ki smo ju uporabljali v diplomski nalogi. Komplet vsebuje robota, kamero, vir osvetlitve, inštalacijsko opremo ter dva programa. V programskem delu najprej opravimo postopek kalibracije in nato sestavimo program za delovanje sistema. Sistem deluje tako, da najprej zajame sliko in jo obdela, nato pa robot izvede strego kock. V zaključku naloge so predstavljene ugotovitve o delovanju sistema, kot so natančnost zaznavanja barve, ponovljivost robota ter njegova primernost v avtomatiziranih procesih.

Keywords

diplomske naloge;strojni vid;robotika;strega materiala;zaznavanje lokacije in barve;nevronske mreže;Bayerjev filter;

Data

Language:	Slovenian
Year of publishing:	2022
Typology:	2.11 - Undergraduate Thesis
Organization:	UL FS - Faculty of Mechanical Engineering
Publisher:	[A. Mugerli]
UDC:	007.52:004.8:004.946(043.2)
COBISS:	134941443
Views:	238
Downloads:	45
Average score:	0 (0 votes)
Metadata:

Other data

Secondary language:	English
Secondary title:	Colour and product orientation detection with machine vision on automatic handling and assembly
Secondary abstract:	This thesis presents a system for detecting the colour and orientation of a cube using computer vision in a combination with a robotic system. Such systems are called machine vision systems, which allow us to automatically detect components help us and to assemble them efficiently and accurately. The first part of the thesis introduction to machine vision, colour image detection and processing, and robotics. Machine vision describes how machine vision works and the most commonly used detection methods. Then, the operation of the sensor used in a machine vision is presented. Machine vision other physical elements are also presented: image detection, such as lens, aperture and exposure, and get a better understanding of the colour detection system in a digital image. The second part describes the hardware and software from Dobot that was used in my thesis. The kit contains the robot, camera, lighting source, installation equipment and two programs. In the software part, we first perform the calibration procedure, then we build the program to run the system. The system works by first capturing the image and processing it, then the robot serves the cubes. The thesis concludes with a discussion of the system's performance, such as colour detection accuracy, robot repeatability and its suitability in automated processes.
Secondary keywords:	thesis;machine vision;robotic;serve of material;location and colour detection;neural networks;Bayer filter;
Type (COBISS):	Bachelor thesis/paper
Study programme:	0
Embargo end date (OpenAIRE):	1970-01-01
Thesis comment:	Univ. v Ljubljani, Fak. za strojništvo
Pages:	XXII, 48 str.
ID:	16401002