Zaznava objektov s strojnim vidom in manipulacija glede na zaznano stanje

Okorn, Nejc (author), Herakovič, Niko (mentor)

Sodobni industrijski sistemi se soočajo z revolucionarnimi spremembami, ki temeljijo na integraciji najnovejših tehnologij in njihovem povezovanju. Gre za četrto industrijsko revolucijo, ki jo imenujemo tudi industrija 4.0. Ta med drugim omogoča izdelavo več različnih izdelkov naenkrat, za kar pa je potrebna tehnologija, ki je sposobna razpoznavanja predmetov in nadaljnjega odločanja glede na njihovo lokacijo. Za prepoznavo in zaznavanje pozicije se največkrat uporablja strojni vid, s katerim dobi računalnik pregled nad procesom. V zaključni nalogi smo izdelali sistem, ki prepoznava in zaznava pozicijo sestavnih delov ter nato glede na pravila, zapisana v razvitem algoritmu, izvede manipulacijo. Sistem temelji na igri človek ne jezi se, saj ta omogoča naključno razporejanje različnih elementov, podobno kot v realnih sistemih. Robot in njegov nasprotnik izmenjaje mečeta kocko in premikata barvne žetone po igralni plošči. Nad ploščo je nameščena spletna kamera, ki analizira elemente na površini. Analiza je sestavljena iz ločene presoje pozicij žetonov in števila pik na kocki. Podatki se nato prenesejo v računalniški program Matlab, ki z zapisanim algoritmom določi, kam se postavi posamezen žeton. Robotska roka uArm izvrši ukaz programa.

• Language: Slovenian
• Keywords: industrija 4.0, strojni vid, robotska roka uArm, Matlab, kamera, krmilni algoritem
• Work type: Bachelor thesis/paper
• Organization: FS - Faculty of Mechanical Engineering
• Year: 2017
• PID: 20.500.12556/RUL-98912
• Publication date in RUL: 13.12.2017

Secondary language

• Language: English
• Title: Machine vision objects recognition and manipulation according to the detected state

Abstract

Modern industrial systems are facing enormous changes in smart technology integration into systems and its connecting, also referred to as industry 4.0. This enables production of products with different characteristics at the same time. This kind of production is possible with visual technology that gives us information about product’s appearance and its locations. Inspection is usually executed with technology called machine vision that provides computer information about the process. In the thesis we created a system consisting of a robot, a camera and software that recognizes positions of parts and manipulates them based on the rules in algorithm. The robot and its opponent roll dice and move colored tokens according to the Ludo game rules. The game is played on a board, monitored by a camera. The latter provides information about dice roll results and the position of tokens. The information is analyzed with MATLAB program. The program decision is transferred to the robotic arm uArm that executes the program command.

• Keywords: industry 4.0, machine vision, uArm robot, Matlab, camera, control algorithm