Izboljšava sistema strojnega vida za kontrolo prisotnosti komponent v procesu montaže

Žunič, Jože (author), Bračun, Drago (mentor)

Proizvodna linija za sestavljanje izdelkov za avtomobilsko industrijo zahteva 100 % kontrolo prisotnosti komponent, ki se ročno nameščajo na za to predvidena mesta. Potrebno je zagotoviti prisotnost in pravilno orientacijo komponent, da v nadaljevanju ne bi proizvajali kose slabe kakovosti oz. poškodovali elemente proizvodne linije med obratovanjem. V okviru diplomske naloge je analizirano stanje vgrajene vizualne kontrole prisotnosti ohišja in podložke na montažnih delovnih mestih. Izboljšano je delovanje s spremembo prijemala na montažnem mestu podložke. Zaradi spremembe je bilo izvedeno ponovno nastavljanje in strojno učenje pametne kamere Keyence. Na delovnem mestu je razvit eksperimentalni, primerljiv sistem strojnega vida. Izbrana je najboljša postavitev kamer kot tudi osvetlitev komponent. Obdelava slike je narejena s programom RoboRealm. Prikazana je primerjava izboljšanega vgrajenega sistema in eksperimentalnega sistema z rezultati obdelave slik. Ohišje zaznavata oba sistema s podobno zanesljivostjo in lahko bi bila uporabljena na avtomatizirani liniji. Podložko zaznava z večjo zanesljivostjo vgrajeni sistem, vendar tudi ta v vgrajenem stanju ni 100 % zanesljiv. Končne ugotovitve poudarjajo pomembnost pravilne zasnove in nastavitve sistema strojnega vida za doseganje visoke zanesljivosti in učinkovitosti na proizvodni liniji.

• Language: Slovenian
• Keywords: avtomatizacija, kontrola prisotnosti, strojni vid, osvetlitev, kamere, RoboRealm
• Work type: Bachelor thesis/paper
• Organization: FS - Faculty of Mechanical Engineering
• Year: 2024
• PID: 20.500.12556/RUL-160188
• Publication date in RUL: 23.08.2024

Secondary language

• Language: English
• Title: Improvement of the machine vision system to control the presence of components in the assembly process

Abstract

The production line for the assembly of automotive products requires 100 % control of the presence of components that are manually installed in the designated places. It is necessary to ensure the presence and correct orientation of components to avoid producing defective pieces or damaging production line elements during operation. The diploma thesis analyses the current state of visual control of the presence of the housing and washer in assembly workplaces. The performance is improved by changing the gripper at the mounting location of the washer. As a result of the change, a realignment and machine learning of the Keyence smart camera was carried out. A comparable machine vision system was developed at the workplace. The most optimal placement of cameras was selected, as was the illumination of the components. Image processing was done using the RoboRealm software. A comparison of the improved current system and the custom system is shown with the results of the image processing. The housing is detected by both systems with similar reliability and could be used on the automated line. The washer is detected with greater reliability by the current system; although even in its current state, it is still not 100 % reliable. The final results emphasize the importance of the correct design and setup of the machine vision system to achieve high reliability and efficiency on the production line.

• Keywords: automation, presence control, machine vision, illumination, cameras, RoboRealm