Računalni vid u upravljanju rampom i identifikaciji vozila

undergraduate thesis

• Title (english): Computer vision in ramp control and vehicle identification
• Author: Luka Majnarić
• Mentor: Dubravko Majetić (mentor)
• Granter: University of Zagreb; Faculty of Mechanical Engineering and Naval Architecture; Zagreb
• Defense date and country: 2023-09-28, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Mechanical Engineering

Abstract

U ovom radu se razmatra uporaba računalnog vida za očitavanje registarskih pločica vozila s ciljem primjene u sustavima koji koriste rampe. Trenira se model za prepoznavanje registarskih pločica, a tekst se očitava pomoću odvojenog modela. Koristi se prijenosno učenje na najnovijim modelima za detekciju objekata. Opisane su njihove arhitekture i načini na koje razni hiperparametri utječu na točnost prepoznavanja. Nakon izolacije registarske pločice, primjenjuje se niz metoda predobrade slike radi povećanja preciznosti prepoznavanja teksta. Odabir modela uzima u obzir rad u industriji, stoga je prioritet ravnoteža između brzine, točnosti i potrebne računalne snage. Predložena je hardverska implementacija cijelog procesa.

Abstract (english)

This paper considers the use of computer vision for reading vehicle license plates with the aim of application in systems that use ramps. A license plate recognition model is trained, and the text is read using a separate model. It uses transfer learning on state-of-the-art object detection models. Their architectures and the ways in which various hyperparameters affect recognition accuracy are described. After isolating the license plate, several image preprocessing methods are applied to increase the accuracy of text recognition. As the selection of the model considers the industry applications, the priority is always the balance between speed, accuracy, and desired computing power. A hardware implementation of the entire process is also proposed.

• Keywords: Računalni vid; Prijenosno učenje; Detekcija objekata; Optičko prepoznavanje znakova; Predobrada slika
• Keywords (english): Computer vision; Transfer learning; Object detection; Optical character recognition; Image preprocessing
• Language: croatian
• URN:NBN: urn:nbn:hr:235:013293
• Study programme: Title: Mechanical Engineering; Study programme type: university; Study level: undergraduate; Academic / professional title: sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka strojarstva (sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka strojarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Open access
• Created on: 2023-09-27 21:57:18