2026-10-14T21:13:50Z https://www.scidb.cn/oai

10.57760/sciencedb.40467 2026-06-18T10:33:32Z

2026-06-18 A multi-species stomatal microscopic image dataset with oriented bounding box annotations from the │ │ │ Loess Plateau doi:10.57760/sciencedb.40467 en This dataset contains 1,527 stomatal microscopic images from 90 plant species (trees, shrubs, and herbs across 20+ families) collected on the Loess Plateau, China. Samples were prepared using the nail-polish impression method and imaged at 20× magnification (3226×3024 px). Each image is annotated with dual-level oriented bounding boxes (OBB) in YOLO format: guard cells (stoma class) and stomatal pores (pore class). The dataset is split into training (1,111 images), validation (278), and test (138) sets. It supports research in stomatal detection, morphological quantification, and cross-species generalization of deep learning models. stomata; deep learning; YOLO; oriented Bounding Box jian xiao hui Chen Guopeng Niu Furong Wang Youlong Yang Ketong EMBARGO https://creativecommons.org/licenses/by-nc/4.0/ dataset Science Data Bank