Computer Vision Dataset for Detecting Cattle Behaviors in Pasture Environments

This dataset contains RGB images and corresponding YOLOv8-format annotation files for the detection and classification of cattle behaviors in pasture-based environments. It was collected as part of the European Horizon 2020 XGain project and is intended for training and validating object detection models using the YOLOv8 framework. The dataset includes high-resolution images captured with overhead cameras and manual bounding box annotations indicating behaviors such as grazing, lying, and standing. Each image-label pair is UUID-matched and organized into a structured folder format. This resource supports research in computer vision, animal welfare, and precision livestock farming.