Prepoznavanje i praćenje objekata u cestovnom prometu

master's thesis

• Title (english): Recognition and monitoring of road traffic objects
• Author: Marija Vučemilo
• Mentor: Marko Jurčević (mentor)
• Committee member: Marko Jurčević (predsjednik povjerenstva)
• Committee member: Jure Konjevod (član povjerenstva)
• Committee member: Ivan Leniček (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2024-02-26, Croatia

Abstract

Ovaj rad bavi se prepoznavanjem i praćenjem objekata koji su neki od problema računalnog vida. Postoji mnogo algoritama koji se mogu iskoristiti za te zadatke, a jedan od njih je YOLO (engl. You Only Look Once). Posljednja verzija modela YOLO, koja je ujedno najbrža i najtočnija, je YOLOv8. U ovom radu, bilo je potrebno trenirati i validirati model YOLOv8 na slikama i videozapisima iz cestovnog prometa. Prije treniranja, bilo je potrebno stvoriti skup podataka s označenim slikama. Nakon svega toga, implementiran je sustav za praćenje objekata. Zadaća tog sustava je detekcija i praćenje objekata na videozapisu iz cestovnog prometa u stvarnom vremenu. Osim toga, aplikacija sprema podatke o detektiranim objektima u bazu podataka te se u datotečni sustav sprema videozapis za svaki objekt. Također, moguć je pregled prošlih događaja i videozapisa. Za implementaciju sustava iskorišten je programski jezik Python.

Abstract (english)

This work deals with the recognition and tracking of objects, which are some of the challenges in computer vision. There are many algorithms that can be used for these tasks, and one of them is YOLO (You Only Look Once). The latest version of the YOLO model, which is also the fastest and most accurate, is YOLOv8. In this work, it was necessary to train and validate the YOLOv8 model on images and videos from road traffic. Before training, it was necessary to create a dataset with annotated images. After all of that, a real-time object tracking system was implemented. The purpose of this system is to detect and track objects in real-time within a live stream of road traffic. Additionally, the application stores data about detected objects in a database, and a video is saved to the file system for each object. It is also possible to review past events and videos. The implementation of the system utilized the Python programming language.

• Keywords: YOLOv8; detekcija objekata; praćenje objekata; Python; OpenCV; PyQt; PostgreSQL
• Keywords (english): YOLOv8; object detection; object tracking; Python; OpenCV; PyQt; PostgreSQL
• Language: croatian
• URN:NBN: urn:nbn:hr:168:469419
• Study programme: Title: Computing; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka računarstva (magistar/magistra inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2024-03-28 07:56:54