Detekcija igrača u snimkama nogometnih utakmica

undergraduate thesis

• Title (english): Detection of Players in Football Game Video
• Author: Fran Kukec
• Mentor: Zoran Kalafatić (mentor)
• Committee member: Zoran Kalafatić (predsjednik povjerenstva)
• Committee member: Petra Bevandić (član povjerenstva)
• Committee member: Igor Stančin (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2020-09-12, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Rad pokriva detekciju igrača na nogometnoj utakmici korištenjem algoritma oduzimanja pozadine. Objašnjeno je što je algoritam oduzimanja pozadine te kako se koristi. Nabrojali smo rekurzivne i nerekurzivne tehnike odvajanja pozadine od objekata u prvom planu. Spomenuti su i dodatni algoritmi koje je bilo moguće razmatrati u okviru rješavanja ovog problema te smo detaljno razmotrili one korištene u implementaciji: MOG2 i KNN. Rješenje je implementirano u programskom jeziku Python i komentiraju se korištene funkcije. Dokumentirana je izrada maske za detekciju objekata u prvom planu. Za naglašavanje i lakšu detekciju igrača na snimci koristili smo morfološke transformacije. Dobivene rezultate označili smo omeđujućim pravokutnicima. Napravili smo evaluaciju uspješnosti detekcije s pripadnim referentnim okvirima i izračunali smo preciznost i odziv.

Abstract (english)

The subject of this thesis is football player detection using a background subtraction algorithm. The background subtraction algorithm is explained and how we used it in implementing the solution. The recursive and non-recursive techniques in background subtraction have been mentioned, as well as other algorithms which we could have used in the solution. The background subtractors MOG2 and KNN, which we used in the implementation are explained. The program was written in the Python programming language and the functions used in the implementation, both ours and imported, are mentioned and commented on. We have covered making the foreground mask from our initial image set. To make player detection easier and more successful, morphological transformations were used. Then, we marked our detections with predicted bounding boxes. After that, the detection was evaluated using respective ground-truth models for each detection and with this data the mean values of precision and recall were calculated.

• Keywords: računalni vid; oduzimanje pozadine; nogomet; detekcija
• Keywords (english): computer vision; opencv; background subtraction; football; detection; python; mog; mog2; knn
• Language: croatian
• URN:NBN: urn:nbn:hr:168:270216
• 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: 2023-01-14 20:54:21