prikaz prve stranice dokumenta Prepoznavanje gesta za upravljanje dronom korištenjem YOLO algoritma
  Download
PDF 5.37 MB
master's thesis
Prepoznavanje gesta za upravljanje dronom korištenjem YOLO algoritma
Rijeka: University of Rijeka, Faculty of Engineering, 2024. urn:nbn:hr:190:553445
Klepac, Kristijan
University of Rijeka
Faculty of Engineering

Institutional repository: Repository of the University of Rijeka, Faculty of Engineering

Cite this document

Klepac, K. (2024). Prepoznavanje gesta za upravljanje dronom korištenjem YOLO algoritma (Master's thesis). Rijeka: University of Rijeka, Faculty of Engineering. Retrieved from https://urn.nsk.hr/urn:nbn:hr:190:553445

Klepac, Kristijan. "Prepoznavanje gesta za upravljanje dronom korištenjem YOLO algoritma." Master's thesis, University of Rijeka, Faculty of Engineering, 2024. https://urn.nsk.hr/urn:nbn:hr:190:553445

Klepac, Kristijan. "Prepoznavanje gesta za upravljanje dronom korištenjem YOLO algoritma." Master's thesis, University of Rijeka, Faculty of Engineering, 2024. https://urn.nsk.hr/urn:nbn:hr:190:553445

Klepac, K. (2024). 'Prepoznavanje gesta za upravljanje dronom korištenjem YOLO algoritma', Master's thesis, University of Rijeka, Faculty of Engineering, accessed 30 November 2024, https://urn.nsk.hr/urn:nbn:hr:190:553445

Klepac K. Prepoznavanje gesta za upravljanje dronom korištenjem YOLO algoritma [Master's thesis]. Rijeka: University of Rijeka, Faculty of Engineering; 2024 [cited 2024 November 30] Available at: https://urn.nsk.hr/urn:nbn:hr:190:553445

K. Klepac, "Prepoznavanje gesta za upravljanje dronom korištenjem YOLO algoritma", Master's thesis, University of Rijeka, Faculty of Engineering, Rijeka, 2024. Available at: https://urn.nsk.hr/urn:nbn:hr:190:553445

Metadata
TitlePrepoznavanje gesta za upravljanje dronom korištenjem YOLO algoritma
Title (english)Gesture recognition for drone control using YOLO algorithm
AuthorKristijan Klepac
MentorZlatan Car (mentor)
Committee memberSaša Vlahinić (predsjednik povjerenstva)
Committee memberZlatan Car (član povjerenstva)
Committee memberIvan Volarić (član povjerenstva)
GranterUniversity of Rijeka
Faculty of Engineering
Rijeka
Defense date and country2024-07-16, Croatia
Scientific / art field,
discipline and subdiscipline		TECHNICAL SCIENCES
Electrical Engineering
Automation and Robotics
Abstract
U ovom diplomskom radu prezentirana je detekcija objekata primjenom YOLO algoritma. Rad započinje s uvođenjem osnovnih pojmova vezanih za umjetne neuronske mreže te se problematika postepeno nadograđuje na konvolucijske neuronske mreže koje igraju ključnu ulogu kada se govori o detekciji objekata i računalnom vidu općenito. Također, uvode se evaluacijske metrike koje se koriste za procjenu uspješnosti modela. Nakon stvaranja teorijske podloge, rad nastavlja s praktičnom primjenom navedenih alata u obliku YOLOv9 algoritma koji je trenutno jedan od najmodernijih i najbržih pristupa detekciji objekata. YOLOv9 algoritam treniran je da prepozna naredbe dane gestikulacijom ruku i tijela s ciljem upravljanja dronom. Subjekt može izdati naredbu "poleti", "sleti" ili "slijedi" te svaku od njih može dati ili položajem svoga tijela i/ili rukom što ukupno čini 6 različitih klasa koje je potrebno detektirati. Na kraju rada model je ispitan na detekciji u stvarnom vremenu koristeci RTMP protokol.
Abstract (english)
In this master’s thesis, object detection using the YOLO algorithm is presented. Thesis begins with the introduction of basic concepts related to artificial neural networks and is gradually upgraded to convolutional neural networks, which play a key role when talking about object detection and computer vision in general. Also, the evaluation metrics that are used to evaluate the performance of the model are introduced. After creating the theoretical basis, the thesis continues with the practical application of the mentioned tools in the form of the YOLOv9 algorithm, which is currently one of the most modern and fastest approaches to object detection. The YOLOv9 algorithm is trained to recognize commands given by hand and body gestures in order to control the drone. The subject can give the command "take off", "land", or "follow", and each of them can be given either by the position of their body and/or by their hand, which makes a total of 6 different classes that need to be detected. At the end of the thesis, the model was tested on real-time detection using the RTMP protocol.
Keywords
Keywords (english)
Languagecroatian
URN:NBNurn:nbn:hr:190:553445
Study programmeTitle: Graduate University Study of Electrical Engineering
Study programme type: university
Study level: graduate
Academic / professional title: magistar/magistra inženjer/inženjerka elektrotehnike (magistar/magistra inženjer/inženjerka elektrotehnike)
Type of resourceText
File originBorn digital
Access conditionsOpen access
Terms of useLicence Attribution 4.0 International (CC BY 4.0) icon
Created on2024-07-14 12:11:38