Datasets:
GrassLiv-3: A UAV-Based Grassland Livestock Detection Dataset
Dataset Description
GrassLiv-3 is a high-resolution UAV imagery dataset developed for livestock detection in grassland environments. It contains 568 UAV images with 40,811 annotated instances across three livestock categories: cattle, horses, and sheep. The dataset covers grassland environments in Inner Mongolia, China, and presents challenges such as small object sizes, dense distributions, occlusions, and complex background textures.
Dataset Statistics
| Category | Instances |
|---|---|
| Cattle | 4,541 |
| Horse | 1,185 |
| Sheep | 35,085 |
| Total | 40,811 |
Related Publication
DAFS-YOLO: Dense-Aware Feature Enhancement for UAV-Based Grassland Livestock Detection in Dense Sheep Scenes.
- Downloads last month
- 34