Sistem za preverjanje napak izdelkov s strojnim vidom in robotom

Paspalovski, Aleksandar (author), Munih, Marko (mentor)

Raziskava se nanaša na vizualno preverjanje napak izdelkov na daljavo s pomočjo strojnega vida in robota, pri čemer izdelek še vedno vizualno pregleduje človek. Zahtevano je bilo avtomatizirano manipuliranje izdelkov in njihovo kvalitetno ter hitro preverjanje na oddaljeni lokaciji v tujini. Cilj naloge je bil narediti prototipni sistem, ki je sestavljen iz ustrezne osvetlitve, kamere in robota. Slednji zagotavlja manipulacijo, osvetlitev in kamera pa omogočata vizualno preverjanje izdelka. Povezave med komponentami prototipnega sistema zagotavljata robotski in računalniški program. Testirali smo en tip izdelka. Zajetih je bilo sedemnajst slik, pri čemer so bile razvidne in uspešno prikazane vse tipične napake. Ugotovljeno je bilo, da je čas robotske manipulacije daljši od časa, ki ga zdaj potrebuje delavec. Sistem deluje zadovoljivo, vendar so možnosti za izboljšave. Pospešitev je mogoča z uporabo dodatne opreme za zajem več slik naenkrat in z izboljšano komunikacijo med komponentami.

• Language: Slovenian
• Keywords: sistem za preverjanje napak, strojni vid, industrijski robot.
• Work type: Master's thesis/paper
• Organization: FE - Faculty of Electrical Engineering
• Year: 2016
• PID: 20.500.12556/RUL-80200
• Publication date in RUL: 11.02.2016

Secondary language

• Language: English
• Title: System for error checking in products with computer vision and the robot

Abstract

The present research refers to remote visual inspection of faulty parts using machine vision and an industrial robot. We were requested to use an automatized manipulation of products using qualitative and swift remote inspection from an abroad location. Following this guideline, a prototype system has been assembeled. Proper lighting, cameras and an industrial robot allow us to manipulate and visualy inspect the part. Communication between components is enabled by both a robotic and a computer programmed algorithm. Inspection tests were conducted on one product type. Results show seventeen images, which successfully present all of the possible faults, which appear on that part. The required inspection time using the prototype system is still greater than the time of a worker. The system shows great promise and has still room for many improvements. These include usage of better equipement and faster communication among components.

• Keywords: inspection system for faulty parts, machine vision, industrial robot.