Avtomatsko zaznavanje, pobiranje in zlaganje naključno orientiranih predmetov

Rozman, Klemen (author), Herakovič, Niko (mentor)

V proizvodnji stremimo k največji zanesljivosti in ponovljivosti vseh procesov, s čimer dosežemo čim manjše zastoje in s tem zmanjšamo izmet. Osnovni procesi v avtomatizirani liniji v sklopu Industrije 4.0 so zlaganje, razvrščanje in ločevanje različnih izdelkov in sestavnih delov, predvsem zaradi stremenja k mešanim linijam, kjer želimo na eni liniji izdelovati več izdelkov. Za izvedbo takšnega procesa se uporablja robote v kombinaciji s strojnim vidom in primernimi algoritmi za krmiljenje procesa. Namen diplomskega dela je razviti algoritem, ki bo v robotovem delovnem prostoru prepoznaval naključno orientirane oblike in barve ter jih zložil na vnaprej določeno mesto. Za preizkus algoritma smo uporabili preizkuševališče s fiksno USB digitalno kamero za zajemanje pozicije in barve ter z uArm robota za manipulacijo kosov. Preizkuševališče simulira stoječ tekoči trak, na katerem do robota pridejo različni raznobarvni naključno orientirani kosi. S kamero se zajame sliko, ki jo z razvitim algoritmom analiziramo, da dobimo podatke o barvi, orientaciji in poziciji posameznih kosov. Podatke se nato pošlje robotski roki, da predmete pobere in zloži vnaprej določene pozicije.

• Language: Slovenian
• Keywords: roboti, krmiljenje, strojni vid, prepoznavanje oblik, naključno orientiranje
• Work type: Final paper
• Typology: 2.11 - Undergraduate Thesis
• Organization: FS - Faculty of Mechanical Engineering
• Place of publishing: Ljubljana
• Publisher: [K. Rozman]
• Year: 2019
• Number of pages: XIV, [28] f.
• PID: 20.500.12556/RUL-111385
• UDC: 007.52:681.5:658.527(043.2)
• COBISS.SI-ID: 16868635
• Publication date in RUL: 29.09.2019

ROZMAN, Klemen, 2019, Avtomatsko zaznavanje, pobiranje in zlaganje naključno orientiranih predmetov [online]. Bachelor’s thesis. Ljubljana : K. Rozman. [Accessed 1 April 2025]. Retrieved from: https://repozitorij.uni-lj.si/IzpisGradiva.php?lang=eng&id=111385

Secondary language

• Language: English
• Title: Automatic detection, picking and sorting of randomly oriented objects

Abstract

In production, we are striving for the highest reliability and repeatability of all processes, thus achieving as little congestion as possible and thus reducing wastes. One of the basic processes in the automated line within Industry 4.0 is the stacking, sorting and separation of various products and components, mainly due to the striving for mixed lines where we want to produce more products on one line. To carry out such a process, robots are used in combination with machine vision and suitable algorithms for process control. The purpose of this diploma is to develop an algorithm that will recognize randomly shaped shapes and colors in the robot working space, and put them in a predetermined place. To test the algorithm we used a test station with a fixed USB digital camera, for position and color capture, and an uArm robot for manipulating pieces. The test station simulates the standing conveyor belt, in which different colors randomly oriented pieces come to the robot. The camera captures the image, and analyzed it with the developed algorithm to obtain information on the color, orientation and position of individual pieces. The data is then sent to the robot to collect items and fold them in a predetermined position.

• Keywords: robots, control, machine vision, image recognition, random orientation