prikaz prve stranice dokumenta Prepoznavanje prometnih znakova za ograničenje maksimalne brzine kretanja za primjenu u naprednim ADAS algoritmima
  Download
PDF 2.06 MB
master's thesis
Prepoznavanje prometnih znakova za ograničenje maksimalne brzine kretanja za primjenu u naprednim ADAS algoritmima
2020. urn:nbn:hr:200:464953
Strišković, Barbara
Josip Juraj Strossmayer University of Osijek
Faculty of Electrical Engineering, Computer Science and Information Technology Osijek
Department of Communications
Chair of Multimedia Systems and Digital Television

Cite this document

Strišković, B. (2020). Prepoznavanje prometnih znakova za ograničenje maksimalne brzine kretanja za primjenu u naprednim ADAS algoritmima (Master's thesis). Osijek: Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek. Retrieved from https://urn.nsk.hr/urn:nbn:hr:200:464953

Strišković, Barbara. "Prepoznavanje prometnih znakova za ograničenje maksimalne brzine kretanja za primjenu u naprednim ADAS algoritmima." Master's thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, 2020. https://urn.nsk.hr/urn:nbn:hr:200:464953

Strišković, Barbara. "Prepoznavanje prometnih znakova za ograničenje maksimalne brzine kretanja za primjenu u naprednim ADAS algoritmima." Master's thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, 2020. https://urn.nsk.hr/urn:nbn:hr:200:464953

Strišković, B. (2020). 'Prepoznavanje prometnih znakova za ograničenje maksimalne brzine kretanja za primjenu u naprednim ADAS algoritmima', Master's thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, accessed 07 December 2024, https://urn.nsk.hr/urn:nbn:hr:200:464953

Strišković B. Prepoznavanje prometnih znakova za ograničenje maksimalne brzine kretanja za primjenu u naprednim ADAS algoritmima [Master's thesis]. Osijek: Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek; 2020 [cited 2024 December 07] Available at: https://urn.nsk.hr/urn:nbn:hr:200:464953

B. Strišković, "Prepoznavanje prometnih znakova za ograničenje maksimalne brzine kretanja za primjenu u naprednim ADAS algoritmima", Master's thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, Osijek, 2020. Available at: https://urn.nsk.hr/urn:nbn:hr:200:464953

Please login to the repository to save this object to your list.
Metadata
TitlePrepoznavanje prometnih znakova za ograničenje maksimalne brzine kretanja za primjenu u naprednim ADAS algoritmima
Title (english)Recognition of maximal speed limit traffic signs for use in advanced ADAS algorithms
AuthorBarbara Strišković
MentorMario Vranješ (mentor)
Committee memberMario Vranješ (predsjednik povjerenstva)
Committee memberRatko Grbić (član povjerenstva)
Committee memberDenis Vranješ (član povjerenstva)
GranterJosip Juraj Strossmayer University of Osijek
Faculty of Electrical Engineering, Computer Science and Information Technology Osijek
(Department of Communications)
(Chair of Multimedia Systems and Digital Television)
Osijek
Defense date and country2020-09-24, Croatia
Scientific / art field,
discipline and subdiscipline		TECHNICAL SCIENCES
Computing
Artificial Intelligence
Abstract
Zadatak ovog diplomskog rada bio je razviti rješenje za prepoznavanje prometnih znakova ograničenja brzine. Kompletno rješenje bilo je potrebno razviti u programskom jeziku C, izvršiti njegovu optimizaciju i implementirati ga na zadanu ADAS (engl. Advanced driver-assistance systems) razvojnu ploču. Kao uvod u rad, dan je pregled postojećih rješenja navedenog problema te su pojašnjene metode i načini koji realiziraju slične zadatke. Detekcija prometnih znakova ograničenja brzine ostvarena je pomoću algoritma koji dolazi uz Vision SDK razvojno okruženje, dok je ispravno prepoznavanje prometnih znakova ostvareno pomoću konvolucijske neuronske mreže (engl. Convolutional Neural Network, CNN ili ConvNet). S obzirom da je CNN bilo potrebno implementirati na zadanu ADAS razvojnu ploču, prilikom kreiranja mreže bilo je potrebno obratiti pozornost na procesorske mogućnosti zadanog sklopovlja. Stoga, umjesto velikog broja konvolucijskih slojeva, željena CNN sadrži samo jedan konvolucijski sloj s 8 filtara veličine 5x5 elemenata slike i softmax sloj. CNN je kreirana u dva programska jezika, programskom jeziku Python za potrebe treniranja CNN na računalu i u programskom jeziku C za implementaciju CNN na ADAS razvojnu ploču. Rezultati testiranja pokazuju određenu preciznost, međutim ona nije zadovoljavajuća na svim korištenim skupovima testnih podataka. Pretpostavlja se da je rezultat klasifikacije prilično nizak zbog problema nastalog prilikom izdvajanja slike iz videozapisa i promjene rezolucije slike te bi daljnji rad svakako trebao ustanoviti da li je dana pretpostavka ispravna i zašto je problem nastao.
Abstract (english)
The main goal of this thesis was to develope a solution for for recognizing speed limit traffic signs. The complete solution had to be developed in the C programming language, optimized and implemented on the default ADAS (Advanced driver-assist systems) development board. As an introduction to the paper, an overview of the existing solutions to the mentioned problems is given, which are explained methods and ways of realizing similar tasks. Speed limit traffic sign detection is achieved using an algorithm that comes with the Vision SDK development environment, while correct traffic sign recognition is achieved using Convolutional Neural Network (CNN or ConvNet). Since CNN needed to be implemented on a given ADAS development board, when creating the network it was necessary to pay attention to the processing capabilities of the given assembly. Therefore, instead of a large number of convolutional layers, the desired CNN contains only one convolutional layer with 8 filters size of 5x5 pixels and a softmax layer. CNN was created in two programming languages, the Python programming language for CNN training on a computer and the C programming language for implementing CNN on the ADAS development board. The test results show some accuracy, however, it is not satisfactory on all test datasets used. It is assumed that the result of the classifications is quite low due to the problem that arose when extracting images from the video and changing image resolutions. Further work should certainly determine whether the given assumption is corrected and why the problem occurred.
Keywords
Keywords (english)
Languagecroatian
URN:NBNurn:nbn:hr:200:464953
Study programmeTitle: Graduate University Study Programme in Electrical Engineering; specializations in: Communications and Informatics, Power Engineering
Course: Communications and Informatics
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 In copyright icon
Created on2020-11-19 10:04:22