Abstract
In this thesis, the potential of utilizing a millimeter-wave frequency modulated continuous wave radar for object detection in the marine environment is investigated. For this, the basic concepts of machine learning, object detection and radar are explained, after which a literature review is conducted to find relevant approaches from existing literature.
Next, data is collected using the radar and an engineering camera, which is then preprocessed using three different approaches. Using state-of-the-art object detection models from the YOLOv8 and YOLOv5u family, the best approach is determined, following which the best performing model is selected.
Finally, the thesis concludes with a discussion about the results, the methodology used and the challenges encountered are discussed.
The method proposed in this thesis was able to archive comparable performance to a camera- based system.
| Uddannelser | MSc in Business Administration and Data Science, (Kandidatuddannelse) Afsluttende afhandling |
|---|---|
| Sprog | Engelsk |
| Udgivelsesdato | 2024 |
| Antal sider | 84 |