University of Twente Student Theses
Monitoring endangered wildlife utilising computer vision models
Arends, A.J.M. (2025) Monitoring endangered wildlife utilising computer vision models.
|
PDF
1MB |
| Abstract: | This research created a semi-automated image classification pipeline which significantly reduces the manual labor required for image classification, with a specific focus on supporting the Boeren, Burgers en Buitenbeesten (BB&B) project. To address this challenge, a systematic literature review was conducted to evaluate current state-of-the-art methods and identify best practices in automated camera trap image classification systems and models. These insights guided the design of a flexible workflow in a Jupyter Notebook setup and led to the selection and fine-tuning of four different candidate models. Among these, a customized ConvNeXt model from Schneider et al. [1], retrained on a limited dataset of mainly the BB&B own dataset, achieved the highest accuracy of 95.45 percent across 22 classes, underscoring the effectiveness of the model. This outcome confirms the successful implementation of the model, which can be utilised inside of the semi-automated classification pipeline for reducing classification effort. |
| Item Type: | Essay (Bachelor) |
| Clients: | Saxion, Enschede, The Netherlands |
| Faculty: | EEMCS: Electrical Engineering, Mathematics and Computer Science |
| Subject: | 43 environmental science, 54 computer science |
| Programme: | Creative Technology BSc (50447) |
| Link to this item: | https://purl.utwente.nl/essays/105242 |
| 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