中国兰品种图像细粒度分类数据集

Orchid2024数据集是一个专为中国兰属兰花（中国兰）品种所设计的细粒度分类数据集。该数据集的样本来源于中国12个省级行政区的20个城市，涵盖8个兰属物种的1269个品种（栽培品种）及6个额外类别，共计156,630张图像。数据集几乎包含了目前中国所有常见的中国兰品种，其细粒度特性与对现实场景的聚焦，使其成为兼具独特性与实用性的科研及实践资源。

The Orchid2024 dataset is a fine-grained classification dataset specifically designed for Chinese Cymbidium (Chinese orchids) cultivars. The samples of this dataset are collected from 20 cities across 12 provincial-level administrative regions in China, covering 1269 cultivars (cultivated varieties) of 8 Cymbidium species plus 6 additional categories, with a total of 156,630 images. This dataset covers almost all currently common Chinese Cymbidium cultivars in China. Its fine-grained nature and focus on real-world scenarios make it a unique and practical resource for both scientific research and practical applications.

搜集汇总

背景与挑战

背景概述

Orchid2024是一个专注于中国兰品种的细粒度图像分类数据集，涵盖8个物种下的1269个品种，共包含156,630张图像，数据来自中国多个地区的真实场景，并呈现长尾分布。该数据集以6:2:2的比例划分为训练、验证和测试集，支持物种和品种两级分类，适用于研究和实践中的品种级识别任务。