magistrsko delo
Tomaž Kavaš (Author), Janez Perš (Mentor)

Abstract

Dandanes se na več področjih uporabljajo sistemi z večimi kamerami. Eno od teh področij je šport. V športu nam sistem zagotavlja pokritje celotnega igrišča. Tako zajamemo obsežno gibanje igralcev skozi celotno igrišče (kot na primer napad v košarki, kjer se lahko celotna ekipa v zelo kratkem času pomakne skozi celotno igrišče). Te informacije so zelo zaželene v športu, saj se uporabijo za analizo igre in za odkrivanje šibkih točk neke ekipe. Vendar pa pridobitev teh informacij ni trivialno. Začetna težava je lahko pri sami sinhronizaciji sistema, saj so v večini primerov kamere prosto tekoče, brez sinhronizacijskega signala. Prav tako pa potrebujemo algoritem, ki lahko iz takega sistema izlušči informacije, ki jih potrebujemo. V tem delu predstavljamo metodo, ki omogoča sinhronizacijo sistema večih kamer. Metoda tudi vključuje algoritem, ki lahko iz sinhroniziranega sistema večih kamer, pridobi 3D točke igralcev in njihove trajektorije. V nalogi smo najprej interpolirali pridobljene posnetke, kar nam je omogočilo sinhronizacijo posnetkov. Po sinhronizaciji smo nato z uporabo nevronske mreže OpenPose na vseh posnetkih detektirali igralce in pridobili njihove 2D točke skeletov. Pridobljene podatke smo nato posredovali sledilcu, ki določi 3D točke igralcev in njihove trajektorije. Na koncu smo dobljene rezultate primerjali z rezultati, ki so bili pridobljeni s tretjim sledilcem in tako ovrednotili naš algoritem.

Keywords

računalniški vid;sledenje;trajektorije;sinhronizacija;sistem večih kamer;magisteriji;

Data

Language: Slovenian
Year of publishing:
Typology: 2.09 - Master's Thesis
Organization: UL FE - Faculty of Electrical Engineering
Publisher: [T. Kavaš]
UDC: 004.8:796(043.3)
COBISS: 79152387 Link will open in a new window
Views: 195
Downloads: 38
Average score: 0 (0 votes)
Metadata: JSON JSON-RDF JSON-LD TURTLE N-TRIPLES XML RDFA MICRODATA DC-XML DC-RDF RDF

Other data

Secondary language: English
Secondary title: Localization of Athletes in Team Sports From Multiple Viewpoints Using Pose Estimation
Secondary abstract: Today multi-camera systems are used in large variety of fields. One of the fields is sports. In sports, the system is used to cover the whole playground. That way we can capture large scale motion during a game (for example, the motion of a whole basketball team running across the entire playground, when performing an attack). All these information from the videos are very desirable in sports, because they can be used to analyze the game and to discover a team's weaknesses. But getting this information is not trivial. The first issue that we could run into is with the synchronization of the whole system, because in most cases free running cameras are used without a synchronizing signal. We also need an algorithm that can extract the information we need from such a system. In this work we present a method that allows us to synchronize a multi-camera system. The method also includes an algorithm that can calculate 3D points of players and their trajectories from a synchronized multi-camera system. We first interpolated the obtained videos, which allowed us to synchronize them. After the synchronization, we used a deep neural network OpenPose to detect all the players on the synchronized videos and get their 2D skeletal points. The obtained data is then used as an input for our algorithm (tracker) to calculate 3D points of the players and their trajectories. Finally we compared our results with the results obtained by another tracker and based on that evaluated the performance of our algorithm.
Secondary keywords: computer vision;tracking;trajectories;sinhronization;multi-camera system;
Type (COBISS): Master's thesis/paper
Study programme: 1000316
Embargo end date (OpenAIRE): 1970-01-01
Thesis comment: Univ. v Ljubljani, Fak. za elektrotehniko
Pages: XVI, 68 str.
ID: 13595018