five

FAIRHiveFrames-1K: A Public FAIR Dataset of 1,265 AnnotatedHive Frame Images with Preliminary YOLOv8/YOLOv11Baselines

收藏
DataCite Commons2026-05-07 更新2026-05-07 收录
下载链接:
https://zenodo.org/doi/10.5281/zenodo.19241078
下载链接
链接失效反馈
官方服务:
资源简介:
FAIRHiveFrames-1K is a publicly available dataset of 1,265 manually annotated hive frame images designed to support research in precision apiculture, computer vision, and AI-enabled agricultural monitoring systems (Agriculture 4.0). The following journal article provides the scientific metadata for FAIRHiveFrames-1K, documents our image captureand annotation methods and FAIR alignment, gives the details of tuning YOLOv8 and YOLOv11 on FAIRHiveFrames-1K to generate preliminary conservative baselines.  Kulyukin, V., Hill, R. and Kulyukin, A. (2026) FAIRHiveFrames-1K: A Public FAIR Dataset of 1265 Annotated Hive Frame Images with Preliminary YOLOv8 and YOLOv11 Baselines. Sensors, 26 (8), 2518.  https://doi.org/10.3390/s26082518 The dataset consists of 1920×1080 PNG images derived from a longitudinal USDA–ARS multi-sensor research reservoir (2013–2022). Each image is annotated with bounding boxes for seven biologically meaningful categories: - CappedHoneyCell- CappedWorkerBroodCell- EmptyCombCell- PollenCell- UncappedNectarCell- UncappedWorkerLarvaCell- BeeHiveFrame In total, the dataset contains 124,669 annotated regions of interest. Annotation was performed using a consensus-based protocol involving two annotators and domain-informed verification. FAIRHiveFrames-1K is distributed with:- PNG images (lossless format)- YOLO plaintext annotations- PascalVOC XML annotations- Fixed train/validation splits- Training and tuning scripts- Pretrained YOLOv8 and YOLOv11 baseline weights The dataset is designed to support reproducible benchmarking under realistic hardware constraints and to facilitate further research in dense small-object detection. DATA ACCESS Due to file size constraints, the full dataset is hosted on Utah State University Boxand can be accessed at: https://usu.box.com/shared/static/kzmjdnzesdf4ssg5zpo5w6xiyycjz7y3.zip This Zenodo record serves as a DOI-registered metadata entry for citation, discovery,and FAIR compliance. For file integrity verification, the SHA256 checksum of the dataset archiveMDPI_Sensors_FAIRHiveFrames_1K.zip is: 280ad37fda41a90d19b79b9339da49504685d83c3f57a7c2f7b8d44578e0ef15. VERSION This record corresponds to: FAIRHiveFrames-1K Version 1.0 (frozen December 31, 2025) Future dataset expansions will follow a size-based versioning scheme(e.g., FAIRHiveFrames-2K, FAIRHiveFrames-5K). SOURCE The dataset is derived from the USDA-ARS Tucson, Arizona research reservoir of multi-year field experiments on managed honey bee colonies.  LICENSE This dataset is distributed under the Creative Commons Attribution 4.0 International (CC BY 4.0) license.
提供机构:
Zenodo
创建时间:
2026-03-26
二维码
社区交流群
二维码
科研交流群
商业服务