COMPUTER GAME CONTROL USING MOTION TRACKING

V diplomskem delu se ukvarjamo z razvojem naravnega vmesnika na osnovi računalniškega vida, s pomočjo katerega uporabnik lahko krmili preprosto računalniško igro. V tej diplomi dejansko preučujemo metode za sledenje gibanju v digitalnih slikah, zajetih s spletno kamero, ki je nameščena na stropu sobe. Osnovni cilj našega algoritma je v realnem času zaznati gibe rok osebe na zajeti digitalni sliki in jih pretvoriti v ukaze, s katerimi krmilimo računalniško igro. Algoritem temelji na preprostih, a časovno zelo učinkovitih metodah računalniškega vida. Naša skoraj realnočasovna rešitev je implementirana v programskem jeziku C# s pomočjo knjižnic OpenCV in Emgu CV. Našo aplikacijo z integriranim naravnim vmesnikom je preizkusilo več testnih oseb, ki so ocenjevale predvsem uporabnost naše rešitve pri krmiljenju spletnih iger. Rezultati so pokazali, da je naš naravni vmesnik primeren za krmiljenje računalniških iger, ki zahtevajo do dva različna ukaza.The diploma deals with the development of a natural interface based on computer vision through which the user can control a simple computer game. More specifically, it examines motion tracking methods for digital images captured with a webcam that is installed on the ceiling of the room. The main goal of our algorithm is real-time hand movement detection on the captured digital image and its conversion into commands that control a computer game. The algorithm is based on simple, but very time-efficient computer vision methods. Our near real-time solution is implemented in C# using OpenCV libraries and Emgu CV. Our application with an integrated natural interface was tested by several test users that evaluated the usefulness of our solution primarily for controlling online games. The results showed our natural control interface to be suitable for computer games that require up to two different commands