Analiza slike završene partije igre Go

undergraduate thesis

• Title (english): Image Analysis of Completed Go Game
• Author: Marko Meštrović
• Mentor: Sven Lončarić (mentor)
• Committee member: Sven Lončarić (član povjerenstva)
• Committee member: Nikola Banić (član povjerenstva)
• Committee member: Ivana Zadro (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2018-09-07, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Electrical Engineering

Abstract

Cilj ovoga završnog rada je bio napraviti program koji će pomoću analize slike olakšati brojanje bodova na kraju partije Go.
Prvo smo filtrirali boje. Pojačavanje kontrasta i binarizacija slike, te filtriranje po predodređenoj funkciji su se pokazale neuspješnim. Filtriranje pomoću Hessiana se pokazalo kao najbolji izbor te je ostatak rada baziran na slikama dobivenima pomoću Hessiana.
Za detekciju linija se koristila Houghova detekcija te se pomoću kutova moglo filtrirati samo korisne linije. Nakon toga se, pomoću udaljenosti od ruba slike, moglo detektirati rubne linije, a zatim i kutove ploče.
Korištenjem funkcija iz Matlab-a, transformirali smo nepravilni četverokut u zadovoljavajući pravokutni oblik. To je omogućilo analitičko određivanje mreže, te se pomoću vrijednosti okolnih piksela svakog raskrižja moglo odrediti da li se na njemu nalazi praznina ili kamenčić, i koje je kamenčić boje.
Brojanje bodova je rješeno tako da su se u prošlom koraku određene vrijednosti spremale u matricu sa 19 redaka i stupaca, te se pomoću okolnih vrijednosti svakoga poena moglo odrediti kojem igraču pripada.

Abstract (english)

The goal of this thesis was to create a program that would help with counting of points at the end of a Go game by using image analysis.
We first filtered the colors. Improving contrast and binarizing the image, followed by filtering with a predefined function proved to be unsuccessful. Filtering using Hessian proved to be the best choice, so the rest of the work was based on images obtained with Hessian.
Hough's detection was used to detect the lines which were then filtered by their angles to get the useful ones. After that, the edge lines were detected by using the distance from the edge of the image, and so were the edges of the board.
Using MATLAB functions we transformed an irregular shape we got from the picture into a rectangular one. This enabled the analytic drawing of the cluster of the Go board. By doing so we got the coordinates of each of the points, and stones, using the median pixel value of the surrounding pixels.
Counting points was done by creating a matrix with 19 rows and columns, and then putting the values of each coordinate into it.

• Keywords: Go; filtriranje boja; detekcija linija; rubne točke; analitički određena mreža
• Keywords (english): Go; filtering the colors; detection of lines; edge points; analytically determined cluster
• Language: croatian
• URN:NBN: urn:nbn:hr:168:265817
• Study programme: Title: Electrical Engineering and Information Technology; Study programme type: university; Study level: undergraduate; Academic / professional title: sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka elektrotehnike i informacijske tehnologije (sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka elektrotehnike i informacijske tehnologije)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2019-03-13 22:59:06