Samodejno prepoznavanje prehodov za pešce na digitalni sliki : diplomska naloga

Dougan, Jernej Nejc (author), Kosmatin Fras, Mojca (mentor), Grigillo, Dejan (co-mentor)

Količine podatkov, ki jih je potrebno obdelati, so vedno večje, zato je smiselno procese avtomatizirati. Digitalne fotografije oziroma slike so vir informacij, ki jih lahko z računalniškimi algoritmi izluščimo in interpretiramo. V diplomski nalogi smo na kratko opisali, kaj je digitalna slika in kako jo lahko obdelujemo. Poudarek naloge je bil na praktičnem delu, kjer smo razvili algoritem za samodejno prepoznavanje prehodov za pešce na digitalni sliki. Program je napisan v programskem okolju Matlab.

• Language: Slovenian
• Keywords: geodezija, diplomska dela, UNI, GIG, prehod za pešce, prepoznavanje vzorcev, korelacija, Houghova transformacija, projektivna transformacija
• Work type: Bachelor thesis/paper
• Typology: 2.11 - Undergraduate Thesis
• Organization: FGG - Faculty of Civil and Geodetic Engineering
• Place of publishing: Ljubljana
• Publisher: [J. N. Dougan]
• Year: 2012
• Number of pages: VIII, 20 str., 6 pril.
• PID: 20.500.12556/RUL-32645
• UDC: 528.7:625.7(043.2)
• COBISS.SI-ID: 6003041
• Publication date in RUL: 10.07.2015

Secondary language

• Language: English
• Title: Automatic pedestrian crossing detection on digital image

Abstract

We have to process more and more data every day, so it makes sense to automate as many processes as possible. Digital images are source of information, which can be extracted and interpreted with computer algorithms. In bachelor thesis we are dealing with digital image and image processing. The emphasis of the thesis is the practical case study. We had developed algorithm for automatic pedestrian crossing detection in digital image. Program is written in Matlab.

• Keywords: graduation thesis, geodesy, geodetic plan, pedestrian crossing, pattern recognition, correlation, Hough transformation, projective transformation