This dataset is used in the bachlors thesis exploring federated learning on drone imagery: "Dean Hoffmann. Federated Learning of Large Language Models on Agricultural Drone Imagery Using Simulated Flight Planning, 2025."
The dataset consists of animal images catured by a drone. Animals captured are: birds (chicken and ducks), cows, horse, and sheep.
The dataset consists of multiple splits. One split contains all images as the drone captured them. Another split contains images in which the animals were cropped. Another split contains the cropped images in different train/val/test splits.
The sheep images orignate from a prior bachlor thesis on a simlar topic: "Mulham Al Maarawi. Exploring chatgpt for drone applications: Prompts, dialogues, and task adaptability. Bachelors thesis, 2024."