| _version_ | 1866901147383496704 |
|---|---|
| author | Amran, Amanda Aiza Chin, Kim On Simon, Donna |
| author_facet | Amran, Amanda Aiza Chin, Kim On Simon, Donna |
| contents | <p>The dataset comprises both tabular and image-based data derived from UAV aerial surveys conducted over Sabah, Malaysia in collaboration with the World Wide Fund for Nature (WWF) Sabah. The image data include raw, processed, annotated (cropped), segmented, and enhanced segmented forms. All images were standardized to 100×100 pixels, focusing specifically on Bornean orangutan nests and non-nest structures such as branches. The raw dataset consists of original UAV-captured imagery, while the processed datasets include images subjected to preprocessing techniques such as background removal using GrabCut, HSV filtering, and bilateral filtering. The segmented and enhanced segmented datasets were generated to improve feature visibility and classification performance, ensuring only visually identifiable nests and branches were included after manual verification.</p> <p>Data collection was carried out using UAV aerial photography across forested regions in Sabah, Malaysia under WWF orangutan nest survey initiatives. All nests and branches were manually cropped into uniform image patches to support consistent machine learning analysis. The dataset was curated to ensure quality control by retaining only images with clear visual distinction between classes.</p> <p>The data originate from Sabah, Malaysia, and are stored at the Faculty of Computing and Informatics, Universiti Malaysia Sabah. The dataset is publicly accessible through the <span class="hover:entity-accent entity-underline inline cursor-pointer align-baseline"><span class="whitespace-normal">Zenodo</span></span> repository under the identifier 17905405, </p> |
| format | Recurso digital |
| id | zenodo_https___doi_org_10_5281_zenodo_17905405 |
| institution | Zenodo |
| language | eng |
| publishDate | 2025 |
| publisher | Zenodo |
| record_format | zenodo |
| spellingShingle | Bornean Orangutan Nests and Branches Classification Dataset Amran, Amanda Aiza Chin, Kim On Simon, Donna <p>The dataset comprises both tabular and image-based data derived from UAV aerial surveys conducted over Sabah, Malaysia in collaboration with the World Wide Fund for Nature (WWF) Sabah. The image data include raw, processed, annotated (cropped), segmented, and enhanced segmented forms. All images were standardized to 100×100 pixels, focusing specifically on Bornean orangutan nests and non-nest structures such as branches. The raw dataset consists of original UAV-captured imagery, while the processed datasets include images subjected to preprocessing techniques such as background removal using GrabCut, HSV filtering, and bilateral filtering. The segmented and enhanced segmented datasets were generated to improve feature visibility and classification performance, ensuring only visually identifiable nests and branches were included after manual verification.</p> <p>Data collection was carried out using UAV aerial photography across forested regions in Sabah, Malaysia under WWF orangutan nest survey initiatives. All nests and branches were manually cropped into uniform image patches to support consistent machine learning analysis. The dataset was curated to ensure quality control by retaining only images with clear visual distinction between classes.</p> <p>The data originate from Sabah, Malaysia, and are stored at the Faculty of Computing and Informatics, Universiti Malaysia Sabah. The dataset is publicly accessible through the <span class="hover:entity-accent entity-underline inline cursor-pointer align-baseline"><span class="whitespace-normal">Zenodo</span></span> repository under the identifier 17905405, </p> |
| title | Bornean Orangutan Nests and Branches Classification Dataset |
| url | https://doi.org/10.5281/zenodo.17905405 |