Metoda analiza video zapisa za brojanje vozila

undergraduate thesis

• Title (english): Video analysis method for vehicle counting
• Author: Mihovil Vinković
• Mentor: Sven Lončarić (mentor)
• Committee member: Sven Lončarić (predsjednik povjerenstva)
• Committee member: Tomislav Petković (član povjerenstva)
• Committee member: Pavle Prentašić (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2014-07-14, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

U radu je opisan postupak brojanja vozila gdje se pozadina modelirala metodom Stauffera i Grimsona (Stauffer, 1999.) preko mješovitoga skupa Gaussovih razdioba. Metoda je detaljno izložena kroz formule te su predložene vrijednosti parametara za opće korištenje algoritma. Algoritam prvo prolazi kroz fazu učenja gdje dobiva inicijalnu pozadinu te onda, u fazi detekcije, kontinuirano detektira te ažurira pozadinu. Zatim dolazi faza analize povezanih komponenti gdje algoritam također i broji vozila preko pamćenja njihovih centroida iz prošloga koraka. Drugi i treći korak se ponavljaju do kraja video-zapisa ili isključenja programa. Postupak je implementiran u okruženju Matlab 2013 uz pomoć nekih gotovih metoda te evaluiran na oko 60 sati testnih snimki. Rezultati su prikazani te komentirani.

Abstract (english)

This paper describes a technique for vehicle counting where the background was modeled using the method described by Stauffer and Grimson (Stauffer, 1999.) where it is modeled using Gaussian mixture models. Method is described in detail through formulae and proposed values of parameters for usage of algorithm. Algorithm first goes through a training phase where it gets an initial background and then, in detection phase, it continuously detects and updates the background. Then comes the connected component analysis phase where the algorithm also counts the vehicles through storing their centroids from the last frame. Second and third steps are repeated until the end of the video-file or program turning off. The algorithm was implemented in environment Matlab 2013 with help of some in-built methods and was evaluated on about 60 hours of test footage. Results are shown and commented on.

• Keywords: subtrakcija pozadine; mješoviti skup Gaussovih razdioba; analiza spojenih komponenti
• Keywords (english): background subtraction; Gaussian mixture models; connected component analysis
• Language: croatian
• URN:NBN: urn:nbn:hr:168:036014
• Study programme: Title: Computing; Study programme type: university; Study level: undergraduate; Academic / professional title: sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka računarstva (sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2018-02-06 07:32:03