Algoritmi za prepoznavanje pokreta i primjena na automatsko praćenje objekata

master's thesis

• Author: Alan Kunštek
• Mentor: Goran Igaly (mentor)
• Committee member: Goran Igaly (predsjednik povjerenstva)
• Committee member: Luka Grubišić (član povjerenstva)
• Committee member: Boris Guljaš (član povjerenstva)
• Committee member: Maja Starčević (član povjerenstva)
• Granter: University of Zagreb; Faculty of Science; (Department of Mathematics); Zagreb
• Defense date and country: 2016-02-24, Croatia
• Scientific / art field, discipline and subdiscipline: NATURAL SCIENCES; Mathematics

Abstract

U ovom radu upoznajemo čitatelja s osnovama područja detekcije pokreta. Ovo područje spada pod računalni vid koji je jedna od gorućih grana računarstva današnjice. Postoje brojne primjene računalnog vida u robotici, automatizaciji, računalnom sučelju, te mnogim drugima. Izabiru se metode koje su prilično lako shvatljive akademskom čitatelju koji nije imao prijašnjih susreta s područjem, te se spominje jedna kompliciranija radi kontrasta. Govori se o glavnim preprekama i izazovima koje potencijalne metode moraju savladati da bi bile upotrebljive za praktičnu primjenu. Kao zaključak rada navodi se prijedlog jednostavne metode za implementaciju u specifičnim kontroliranim uvjetima. Time se pokazuje da unatoč tome što u području postoji na stotine pristupa, metoda i varijacija ne postoji jedna najbolja za sve primjene. Glavna nit vodilja kod odabira metode ipak na kraju mora biti minimizacija zahtjeva na računalne resurse, a da metoda ipak daje iskoristive rezultate.

Abstract (english)

In this thesis we introduce the reader to the basics of motion detection. Motion detection is part of computer vision, which is one of the thriving branches of contemporary computer science. There are many applications for computer vision in robotics, automatization, computer interface, and many others. We choose methods which are easily understandable to academic reader with no previous experience in the field, and we mention one which is a bit more complicated, for contrast. We talk about main obstacles and challenges which potential methods must overcome in order to be viable for application. As a conclusion to thesis we propose a simple method for implementation in controlled conditions. Thus, we show that despite the fact that motion detection has hundreds of approaches, methods and their variations, there isn’t a single one that is best for all applications. In general, as we chose a method to use in specific case, we must try to minimize computer resource consumption, while keeping the method results usable for intended purpose.

• Keywords: detekcija pokreta; računalni vid
• Keywords (english): motion detection; computer vision
• Language: croatian
• URN:NBN: urn:nbn:hr:217:519947
• Study programme: Title: Computer Science and Mathematics; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra računarstva i matematike (magistar/magistra računarstva i matematike)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Open access
• Created on: 2019-02-20 13:24:47