prikaz prve stranice dokumenta Automatska detekcija i brojanje automobila na video snimkama
Access restricted to authenticated users
master's thesis
Automatska detekcija i brojanje automobila na video snimkama
Split: University of Split, Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture, 2022. urn:nbn:hr:179:558596
Jerčić, Borna
University of Split
Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture
Department of Electronics and Computing

Institutional repository: Repository of the Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture

Cite this document

Jerčić, B. (2022). Automatska detekcija i brojanje automobila na video snimkama (Master's thesis). Split: University of Split, Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture. Retrieved from https://urn.nsk.hr/urn:nbn:hr:179:558596

Jerčić, Borna. "Automatska detekcija i brojanje automobila na video snimkama." Master's thesis, University of Split, Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture, 2022. https://urn.nsk.hr/urn:nbn:hr:179:558596

Jerčić, Borna. "Automatska detekcija i brojanje automobila na video snimkama." Master's thesis, University of Split, Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture, 2022. https://urn.nsk.hr/urn:nbn:hr:179:558596

Jerčić, B. (2022). 'Automatska detekcija i brojanje automobila na video snimkama', Master's thesis, University of Split, Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture, accessed 11 July 2024, https://urn.nsk.hr/urn:nbn:hr:179:558596

Jerčić B. Automatska detekcija i brojanje automobila na video snimkama [Master's thesis]. Split: University of Split, Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture; 2022 [cited 2024 July 11] Available at: https://urn.nsk.hr/urn:nbn:hr:179:558596

B. Jerčić, "Automatska detekcija i brojanje automobila na video snimkama", Master's thesis, University of Split, Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture, Split, 2022. Available at: https://urn.nsk.hr/urn:nbn:hr:179:558596

Metadata
TitleAutomatska detekcija i brojanje automobila na video snimkama
Title (english)Automatic detection and counting of cars on video
AuthorBorna Jerčić
MentorMaja Braović https://orcid.org/0000-0002-3655-3534 (mentor)
GranterUniversity of Split
Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture
(Department of Electronics and Computing)
Split
Defense date and country2022-06-30, Croatia
Scientific / art field,
discipline and subdiscipline		TECHNICAL SCIENCES
Computing
Abstract
U ovom radu opisane su konvolucijske neuronske mreže, njihova osnovna arhitektura i prezentirane arhitekture postojećih konvolucijskih neuronskih mreža. Opisane su korištene tehnologije za implementaciju programa, a to su: Python, OpenCV, Tensorflow i Keras. Također su opisani postojeći algoritmi automatske detekcije objekata te implementirana i testirana vlastita metoda temeljena na tradicionalnim metodama računalnog vida kao i na modernim metodama dubokog učenja, tj. konvolucijskih neuronskih mreža. Prikupljen je vlastiti skup podataka za treniranje i validaciju podijeljen u šest kategorija najčešćih objekata na prometnim snimkama različitih cesta i ulica i jednu kategoriju za detekciju okoline i praznih cesta te na temelju tih podataka izrađen model konvolucijske neuronske mreže za klasifikaciju koji osim klasifikacije objekata ima i ulogu odbacivanje krivih detekcija. Dana je usporedba postojećih algoritama s implementiranim, napravljena analiza brzina detekcije te dane ideje o daljnjem razvoju programa.
Abstract (english)
This paper describes convolutional neural networks, their basic architecture and presents the architectures of existing convolutional neural networks. The technologies used for the implementation of the program are described, namely: Python, OpenCV, Tensorflow and Keras. The existing algorithms for automatic object detection are also described, as well as my own implemented and tested method based on traditional methods of computer vision and modern methods of deep learning, ie convolutional neural networks. A personal set of data for training and validation divided into six categories of the most common objects on traffic videos of different roads and streets and one category for detection of environment and empty roads for rejection of false detections was collected. Based on these data a convolutional neural network model for classification was created. The comparison of existing algorithms with the implemented one and the analysis of detection speeds is made. Ideas on further development of the program are given.
Keywords
Keywords (english)
Languagecroatian
URN:NBNurn:nbn:hr:179:558596
Study programmeTitle: Computing
Study programme type: university
Study level: graduate
Academic / professional title: magistar/magistra inženjer/inženjerka računarstva (magistar/magistra inženjer/inženjerka računarstva)
Type of resourceText
File originBorn digital
Access conditionsAccess restricted to authenticated users
Terms of useLicence In copyright icon
Created on2023-05-18 08:22:36