Detekcija semafora zasnovana na analizi svijetlih područja slike

master's thesis

• Title (english): Traffic light detection based on blob detection
• Author: Vedrana Raguž
• Mentor: Sonja Grgić (mentor)
• Committee member: Sonja Grgić (predsjednik povjerenstva)
• Committee member: Lea Skorin-Kapov (član povjerenstva)
• Committee member: Dario Bojanjac (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2020-07-14, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Electrical Engineering

Abstract

Sustavi za pomoć vozaču pružaju informacije o stanju prometa. Detekcija semafora je bitan element osiguranja sigurnosti za vrijeme vožnje. U ovom radu je opisan sustav za detekciju semafora zasnovan na analizi svijetlih područja slike. Pri razvoju se koristila Bosch Small Traffic Lights Dataset baza. Dostupni skup uzoraka se proširio metodom horizontalnog preslikavanja. Proučavale su se metode korisne za detekciju semafora poput uklanjanja šuma, metoda postavljanja praga za određivanje komponente boje, detekcije izvora svjetlosti i morfoloških operacija. Konačni algoritam se sastoji od uklanjanja šuma, detekcije izvora svjetlosti i filtriranja područja koja nisu od interesa. Evaluacija sustava je prikazana TP i FP mjerama.

Abstract (english)

Advanced Driver Assistance Systems (ADAS) provide traffic condition information. Traffic light detection is an essential component for ensuring safety while driving. This master thesis describes a traffic light detection system based on blob analysis. For development, Bosch Small Traffic Lights Dataset was used. Available data samples were augmented using horizontal flip method.This thesis also studies methods useful for traffic light detection such as noise reduction, color thresholding, spot light detection and morphological operations. The final algorithm consists of noise reduction, spot light detection and filtering areas that are not of interest. System evaluation is presented in the form of TP and FP measures.

• Keywords: detekcija semafora; svijetlo područje; područje od interesa; HSV; ADAS
• Keywords (english): traffic light detection; blob; region of interest; HSV; ADAS
• Language: croatian
• URN:NBN: urn:nbn:hr:168:327939
• Study programme: Title: Information and Communication Technology; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka informacijske i komunikacijske tehnologije (magistar/magistra inženjer/inženjerka informacijske i komunikacijske tehnologije)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2023-01-13 10:39:41