Li, JiaqiGuala, MicheleHong, Jiarong2022-10-062022-10-062022-10-06https://hdl.handle.net/11299/241878The dataset includes a version of original images and a version of images with augmentation (rotation, exposure, blur, salt & pepper noise). Each version consists of a training set and a validation set, each with the same structure: images and labels. The images are in JPEG format, each with at least 16 snow particles. The labels are in TXT format, including the classes (in numbers: 0-aggregate/irregular/I, 1-dendrite/P2, 2-graupel/rime/R, 3-plate/P1, 4-needle/column/N/C, and 5-small particles/germ/G) assigned to the snow particles, and the bounding boxes (including location coordinates and normalized height/width of the box) for detecting them.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).CC0 1.0 UniversalHolographySnowflakesMachine learningDatasethttps://doi.org/10.13020/7akr-8n50