This readme.txt file was generated on 2022-10-16 by Recommended citation for the data: Li, Jiaqi; Guala, Michele; Hong, Jiarong. (2022). Database of snow holograms collected from 2019 to 2022 for machine learning training or other purposes. Retrieved from the Data Repository for the University of Minnesota, https://doi.org/10.13020/7akr-8n50. ------------------- GENERAL INFORMATION ------------------- 1. Title of Dataset: Database of snow holograms collected from 2019 to 2022 for machine learning training or other purposes 2. Author Information Author Contact: Jiaqi Li (li001334@umn.edu) Name: Jiaqi Li Institution: University of Minnesota Email: li001334@umn.edu ORCID: 0000-0002-1201-7489 Name: Michele Guala Institution: University of Minnesota Email: mguala@umn.edu ORCID: Name: Jiarong Hong Institution: University of Minnesota Email: jhong@umn.edu ORCID: 3. Date published or finalized for release: 2022-10-06 4. Date of data collection (single date, range, approximate date): 2018-01-14 to 2022-04-17 5. Geographic location of data collection (where was data collected?): EOLOS Field Research Station, Rosemount, MN St Anthony Falls Laboratory, Minneapolis, MN 6. Information about funding sources that supported the collection of the data: The work related to this dataset is supported by the National Science Foundation (Program Manager, Nicholas Anderson) under grant NSF-AGS-1822192. 7. Overview of the data (abstract): This dataset includes the original combined snow holograms and holograms with image augmentation (rotation, exposure, blur, noise) for YOLOv5 model training to detect and classify snow particles. The individual snow particles are cropped and combined to enrich the particle numbers in each image for the ease of manual labeling. The snow particles are classified into six categories, including aggregate/irregular (I), dendrite (P2), graupel/rime (R), plate (P1), needle/column (N/C), and small particles/germ (G). -------------------------- SHARING/ACCESS INFORMATION -------------------------- 1. Licenses/restrictions placed on the data: CC0 1.0 Universal (http://creativecommons.org/publicdomain/zero/1.0/) 2. Links to publications that cite or use the data: Li, J.; Guala, M.; Hong, J. Snow Particle Analyzer for Simultaneous Measurements of Snow Density and Morphology. Journal of geophysical research. Atmospheres 2023. https://doi.org/10.1029/2023JD038987 3. Was data derived from another source? If yes, list source(s): dataset augmentation using https://roboflow.com/ 4. Terms of Use: Data Repository for the U of Minnesota (DRUM) By using these files, users agree to the Terms of Use. https://conservancy.umn.edu/pages/drum/policies/#terms-of-use --------------------- DATA & FILE OVERVIEW --------------------- File List A. Filename: Snow_particles_original_yolov5pytorch.zip Short description: Training and Validation Data with original images B. Filename: Snow_particles_augmented_yolov5pytorch.zip Short description: Training and Validation Data with augmented images 2. Relationship between files: List A includes original holograms collected and List B includes holograms after data augmentation (rotate, varying exposure, add noise) -------------------------- METHODOLOGICAL INFORMATION -------------------------- 1. Description of methods used for collection/generation of data: The holograms are collected using the Snow Particle Analyzer based on a digital inline holography system and a high precision scale (detailed description in https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2023JD038987). Here in this dataset, only the holograms from the digital inline holography system are included. Individual snow particles are detected and cropped out from the raw holograms and patched together to form this dataset. 2. Methods for processing the data: In this dataset, we classified all the snow particles and labeled the bounding boxes enclosing them (in the text files in labels). This data is shared for reference in terms of snow particle classification and for the training of potential machine-learning models for snow particle classification. 3. Instrument- or software-specific information needed to interpret the data: This dataset is formatted in a way for the training of object detection and classification models (YOLOv5 to be precise). It can be easily read using the data loader of YOLOv5. 4. Standards and calibration information, if appropriate: The holograms has a spatial resolution of 14.3 um/px. 5. Environmental/experimental conditions: The datasets are collected during multiple deployments at snowy nights at the EOLOS Field Research Station and St. Anthony Falls Laboratory. Details can be found in https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2023JD038987.