Prepoznavanje značajki ronioca u sonarskoj slici

master's thesis

• Title (english): Diver Features Detection in Sonar Images
• Author: Zdravko Čićin-Šain
• Mentor: Nikola Mišković (mentor)
• Committee member: Nikola Mišković (predsjednik povjerenstva)
• Committee member: Ivan Marković (član povjerenstva)
• Committee member: Đula Nađ (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2021-02-26, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Electrical Engineering

Abstract

U ovom diplomskom radu detaljno su opisani postupci potrebni za dobivanje modela za detekciju ronioca u sonarskim slikama. U prvom dijelu rada opisane su vrste anotacija i alati za njihovo generiranje te najbolje prakse kod njihove kreacije. Drugi dio rada započinje opisom izvora korištenih podataka te njihovom pripremom za korištenje s konvolucijskim neuronskim mrežama. Dalje slijedi opis korištenih alata i biblioteka te postupak generiranja i postavke potrebnih konfiguracijskih datoteka za treniranje modela. Na samom kraju opisani su kriteriji evaluacije modela te su isti primijenjeni na dobivenim modelima.

Abstract (english)

In this thesis we gave a detailed description of the procedures required to obtain a model for detecting divers in sonar images. The first part of the paper describes types of annotations and tools for their generation, as well as best practices used in their creation. The second part of the paper begins with a description of the data used and its sources, as well as the process of preparing said data for use with convolutional neural networks. Following is a description of the used tools, libraries and frameworks used in the process of generating and setting up the necessary configuration files for model training. At the very end, the evaluation criteria of the models are described, which have been applied on the obtained models.

• Keywords: Detekcija objekata; Granične kutije; Anotacija linijama; Poligonalna segmentacija; Segmentacija instanci; Anotacija orijentira; 3D kubična anotacija; XML; JSON; YOLO.txt; Darknet biblioteka; Google Colab bilježnica; Jupyter bilježnica; Anaconda
• Keywords (english): Object detection; Bounding Boxes; Line Annotation; Polygonal Segmentation; Semantic Segmentation; Instance Segmentation; Landmark Annotation; 3D Cuboid Annotation; XML; JSON; YOLO.txt; Darknet framework; Google Colab notebook; Jupyter notebook; Anaconda
• Language: croatian
• URN:NBN: urn:nbn:hr:168:333445
• Study programme: Title: Electrical Engineering and Information Technology; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka elektrotehnike i informacijske tehnologije (magistar/magistra inženjer/inženjerka elektrotehnike i informacijske tehnologije)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2023-01-15 09:55:45