Detekcija prometnih traka pomoću konvolucijskih neuronskih mreža

master's thesis

• Title (english): Convolutional Neural Networks for Detection of Traffic Lanes
• Author: Filip Janđel
• Mentor: Vladimir Čeperić (mentor)
• Committee member: Vladimir Čeperić (predsjednik povjerenstva)
• Committee member: Zoran Kalafatić (član povjerenstva)
• Committee member: Marko Koričić (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2021-09-22, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

U radu je predstavljen problem detekcije prometnih traka. Opisane su umjetne neuronske mreže s naglaskom na potpuno konvolucijske neuronske mreže, pomoću kojih
računalo uči detektirati prometne trake na ulaznim slikama. Prikazani su skupovi podataka s kojima mreža radi, te je opisan proces njihovog dohvaćanja pomoću CARLA
programskog okvira za simulaciju kretnje i upravljanja vozila. Implementiran je programski sustav temeljen na modelu umjetne potpuno konvolucijske neuronske mreže.
Prikazani su dobiveni rezultati učenja i testiranja implementirane mreže. Na kraju su
rezultati analizirani, te su komentirani mogući problemi implementacije kao i moguće
buduće nadogradnje modela.

Abstract (english)

In this thesis, the problem of detection of traffic lanes is introduced. Artificial neural networks are described with an emphasis on fully convolutional neural networks
by which computer learns how to detect traffic lanes on input images. Data sets, which
are used by network, are shown together with description of process of gathering data
from CARLA software framework for vehicle driving simulation. A software system based on model of artificial fully convolutional neural network is implemented.
Obtained results for training and testing of the implemented model are shown. Lastly, the results are analyzed together with possible problems of implementation and possible future
extensions of the model.

• Keywords: neuronske mreže; konvolucijske neuronske mreže; učenje; testiranje; segmentacija slike; augmentacija slike; detekcija prometnih traka; skupovi podataka
• Keywords (english): neural networks; convolutional neural networks; training; testing; image segmentation; image augmentation; detection of traffic lanes; data sets
• Language: croatian
• URN:NBN: urn:nbn:hr:168:366645
• 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: 2023-01-13 12:30:46