prikaz prve stranice dokumenta Detekcija i praćenje pješaka u okviru autonomne vožnje
  Download
PDF 3.49 MB
master's thesis
Detekcija i praćenje pješaka u okviru autonomne vožnje
Osijek: Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, 2024. urn:nbn:hr:200:790927
Jusup, Mario
Josip Juraj Strossmayer University of Osijek
Faculty of Electrical Engineering, Computer Science and Information Technology Osijek
Department of Computer Engineering and Automation
Chair of Automation and Robotics

Institutional repository: Faculty of Electrical Engineering, Computer Science and Information Technology Osijek

Cite this document

Jusup, M. (2024). Detekcija i praćenje pješaka u okviru autonomne vožnje (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:200:790927

Jusup, Mario. "Detekcija i praćenje pješaka u okviru autonomne vožnje." Master's thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, 2024. https://urn.nsk.hr/urn:nbn:hr:200:790927

Jusup, Mario. "Detekcija i praćenje pješaka u okviru autonomne vožnje." Master's thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, 2024. https://urn.nsk.hr/urn:nbn:hr:200:790927

Jusup, M. (2024). 'Detekcija i praćenje pješaka u okviru autonomne vožnje', Master's thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, accessed 24 November 2024, https://urn.nsk.hr/urn:nbn:hr:200:790927

Jusup M. Detekcija i praćenje pješaka u okviru autonomne vožnje [Master's thesis]. Osijek: Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek; 2024 [cited 2024 November 24] Available at: https://urn.nsk.hr/urn:nbn:hr:200:790927

M. Jusup, "Detekcija i praćenje pješaka u okviru autonomne vožnje", Master's thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, Osijek, 2024. Available at: https://urn.nsk.hr/urn:nbn:hr:200:790927

Metadata
TitleDetekcija i praćenje pješaka u okviru autonomne vožnje
Title (english)Detection and tracking of pedestrians within the autonomous driving
AuthorMario Jusup
MentorRatko Grbić (mentor)
Committee memberRatko Grbić (predsjednik povjerenstva)
Committee memberDenis Vranješ (član povjerenstva)
Committee memberMario Vranješ (član povjerenstva)
GranterJosip Juraj Strossmayer University of Osijek
Faculty of Electrical Engineering, Computer Science and Information Technology Osijek
(Department of Computer Engineering and Automation)
(Chair of Automation and Robotics)
Osijek
Defense date and country2024-10-01, Croatia
Scientific / art field,
discipline and subdiscipline		TECHNICAL SCIENCES
Computing
Artificial Intelligence
Abstract
U ovom radu opisan je razvoj rješenja za detekciju i praćenje pješaka u okviru autonomne vožnje. Prvo je objašnjen moderan način detekcije i praćenja pješaka u digitalnim slikama. Nakon toga je opisano nekoliko postojećih rješenja za detekciju i praćenje objekata. Rješenje se temelji na YOLOv7 algoritmu za detekciju i SORT i DeepSORT algoritmima za praćenje. Algoritam za detekciju treniran je na vlastitom skupu podataka koji se sastoji od više od sedam tisuća slika. Analizirano je nekoliko dostupnih skupova podataka koji se sastoje od slika i oznaka u obliku graničnih okvira i odabrana su tri najrelevantnija skupa za ovaj problem. Detektor je treniran metodom prijenosnog učenja i metodom bez unaprijed postavljenih težina. Rješenje je potrebno implementirati i na ugradbenu platformu NVIDIA Jetson Nano pa je potrebno koristiti i YOLOv7 tiny verziju detektora s smanjenim brojem parametara. Nakon treniranja razvijen je algoritam koji kombinira algoritme za detekciju i za praćenje pješaka. Korištenjem SORT i DeepSORT algoritama za praćenje povećana je brzina izvođenja predloženog algoritma za detekciju i praćenje pješaka čime je omogućeno izvođenje ovog računalno zahtjevnog algoritma u stvarnome vremenu na računalima s ograničenim resursima. Evaluacija rješenja izvedena je na četiri videozapisa koja simuliraju stvarne uvjete.
Abstract (english)
This paper describes the development of a solution for the detection and tracking of pedestrians in the context of autonomous driving. First, a modern way of detecting and tracking pedestrians in digital images is explained. After that, several existing solutions for object detection and tracking are described. The solution is based on the YOLOv7 detection algorithm and the SORT and DeepSORT tracking algorithms. The detection algorithm was trained on customized data set consisting of more than seven thousand images. Several available datasets consisting of images and labels in the form of bounding boxes were analyzed and the three most relevant datasets for this problem were selected. The detector was trained using a transfer learning method and a method without pretrained weights. The solution needs to be implemented on the embeded NVIDIA Jetson Nano platform, so it is necessary to use the YOLOv7 tiny version of the detector with a reduced number of parameters. After training, an algorithm was developed that combines algorithms for detection and tracking of pedestrians. By using SORT and DeepSORT tracking algorithms, the execution speed of the proposed algorithm for pedestrian detection and tracking has been increased, which enables the execution of this computationally complex algorithm in real time on computers with limited resources. The evaluation of the solution was performed on four videos simulating real world conditions.
Keywords
Keywords (english)
Languagecroatian
URN:NBNurn:nbn:hr:200:790927
Study programmeTitle: Graduate University Study Programme in Computer Engineering
Study programme type: university
Study level: graduate
Academic / professional title: sveučilišni/a magistar/magistra inženjer/inženjerka računarstva (sveučilišni/a magistar/magistra inženjer/inženjerka računarstva)
Type of resourceText
File originBorn digital
Access conditionsOpen access
Terms of useLicence In copyright icon
Created on2024-10-04 11:04:53