This dataset contains YOLOv8-compliant bounding box annotation files in `.txt` format. Each file matches a corresponding RGB image and provides object class labels along with normalized coordinates for detected objects (e.g., cow grazing, cow lying). The labels follow the standard YOLO format: ` `, with all values normalized to the range [0,1]. These annotations were manually curated to enable supervised training of computer vision models in pasture-based livestock settings.