In this thesis, I will present the use of ArUco tag markers in the context of self-driving boats. I will begin by introducing what ArUco markers are and why they are useful for the navigation task. I will then explain camera calibration and describe the procedure to achieve it. Finally, I will discuss the results we obtained from our experiments. In Chapter 2, I will focus on object detection. To achieve this task, I will introduce and explain Grounding DINO for object detection and unfold how it works and how it can be applied and improved to fit the dataset we have. Finally, in Chapter 3, I will discuss the aspects of the project that can still be improved and show some possible directions for future development. I will touch on various elements: from live detection to obstacle classification. I will also reflect on how the work carried out in this thesis may serve as a foundation for further research.
In questa tesi verranno trattati gli ArUco markers nell'ambito delle barche a guida autonoma. Verranno definiti motivandone la scelta progettuale. Al fine di poter utilizzare i markers verrà poi introdotta la calibrazione della telecamera sia dal punto di vista teorico, sia come procedura pratica. Per concludere il capitolo verranno mostrati i risultati ottenuti facendo eseguire l'algoritmo di riconoscimento in due diverse qualità: 480p e 1080p Nel secondo capitolo mi concentrerò sull'object detection mostrando grounding DINO e spiegandone il funzionamento nel dettaglio. Inoltre, verranno mostrati i risultati ottenuti eseguendo l'algoritmo sul nostro dataset. Per concludere, nel terzo capitolo verranno trattati i punti ancora migliorabili: dall'analisi live alla classificazione degli oggetti, mostrando come ogni futura miglioria potrà usare questo progetto come una solida base.
Integrating ArUco Markers and Grounding DINO for Vision-Based Obstacle Detection
PADRIN, MATTEO
2024/2025
Abstract
In this thesis, I will present the use of ArUco tag markers in the context of self-driving boats. I will begin by introducing what ArUco markers are and why they are useful for the navigation task. I will then explain camera calibration and describe the procedure to achieve it. Finally, I will discuss the results we obtained from our experiments. In Chapter 2, I will focus on object detection. To achieve this task, I will introduce and explain Grounding DINO for object detection and unfold how it works and how it can be applied and improved to fit the dataset we have. Finally, in Chapter 3, I will discuss the aspects of the project that can still be improved and show some possible directions for future development. I will touch on various elements: from live detection to obstacle classification. I will also reflect on how the work carried out in this thesis may serve as a foundation for further research.| File | Dimensione | Formato | |
|---|---|---|---|
|
Matteo_Padrin.pdf
accesso aperto
Dimensione
2.19 MB
Formato
Adobe PDF
|
2.19 MB | Adobe PDF | Visualizza/Apri |
The text of this website © Università degli studi di Padova. Full Text are published under a non-exclusive license. Metadata are under a CC0 License
https://hdl.handle.net/20.500.12608/91683