Dataset of Labeled Images for Computer Vision-Based Identification of Dairy Cows for Automated Methane Monitoring

This dataset was generated as part of the project “Development of an Electronic Device for Automated Individual Monitoring of Enteric Methane Emissions in Dairy Cows”. It contains labeled images of 44 individual Holstein dairy cows, captured in a commercial milking parlor using overhead cameras.Each image focuses on the head and neck region of a single cow and is accompanied by a YOLO-format annotation file containing bounding-box information for re-identification purposes. The dataset includes:– 2,121 cropped images– 44 classes (one per cow)– Split into training (1,678 images) and validation (443 images) sets– A configuration file (data.yaml) for direct use with YOLO-based models The resource is intended to support the development and fine-tuning of computer-vision models (e.g., YOLOv8 and YOLOv10) for individual cow identification, enabling the integration of low-cost methane sensors with vision-based animal recognition to attribute emissions to specific animals. Data were collected in a real commercial farm setting and prepared to be readily usable for model training and benchmarking.The dataset is distributed under the CC BY-NC-SA 4.0 license, allowing reuse for non-commercial research and educational purposes with proper attribution.