University of Twente Student Theses
Development and Evaluation of an AI-Driven Pipeline for Wildlife Monitoring
Antonov, Alexandar (2025) Development and Evaluation of an AI-Driven Pipeline for Wildlife Monitoring.
|
PDF
129kB |
| Abstract: | Abstract—Bird populations in the Netherlands have sharply declined in recent years, driven by habitat loss, and increased predation in fragmented landscapes. Camera traps offer a non-intrusive way to monitor wildlife, yet they produce vast numbers of noisy images triggered by non- animal events. This research addresses that challenge by developing and evaluating an AI-driven solution capable of detecting animals in such challenging environments. A data pipeline encompasses end-to-end data preparation and preprocessing that goes from raw camera data to prepared quality data for training. As well as the preparation, deployment, fine-tuning and comparison of the cutting-edge object detection models YOLOv5, DETR and Grounding DINO. This research reports its discovery on YOLOv5s lightweights’ architecture and ability for time sensitive. tasks DETRs transformer advantages and slow convergence drawbacks and Grounding Dinos impressive results and unique qualities at the price of computational and resource overheads. |
| Item Type: | Essay (Bachelor) |
| Faculty: | EEMCS: Electrical Engineering, Mathematics and Computer Science |
| Subject: | 43 environmental science, 50 technical science in general, 54 computer science |
| Programme: | Computer Science BSc (56964) |
| Link to this item: | https://purl.utwente.nl/essays/105261 |
| Export this item as: | BibTeX EndNote HTML Citation Reference Manager |
Show download statistics for this publication
Show download statistics for this publicationRepository Staff Only: item control page