ARBORETUM
收藏资源简介:
ARBORETUM是由爱荷华州立大学等机构创建的大型多模态数据集,旨在推动AI在生物多样性领域的应用。该数据集包含1.346亿张图像,涵盖鸟类、蜘蛛、昆虫、植物、真菌、软体动物和爬行动物等七大类群,总计约32.69万种物种。每张图像均附有科学名称、分类细节和常用名称,由领域专家审核确保准确性。数据集通过iNaturalist社区科学平台收集,经过精心筛选和处理,支持多模态视觉-语言AI模型的训练,特别适用于生物多样性评估和农业研究。ARBORETUM不仅促进了AI模型的发展,还为从害虫控制到全球生物多样性评估和环境保护等数字工具的开发提供了关键支持。
ARBORETUM is a large-scale multimodal dataset developed by institutions including Iowa State University, aiming to advance the application of AI in the field of biodiversity. This dataset contains 134.6 million images, covering seven major taxa such as birds, spiders, insects, plants, fungi, mollusks and reptiles, with approximately 326,900 species in total. Each image is accompanied by scientific names, taxonomic details and common names, which have been verified for accuracy by domain experts. The dataset is collected via the iNaturalist community science platform, and has undergone rigorous screening and processing, supporting the training of multimodal vision-language AI models, and is particularly suitable for biodiversity assessment and agricultural research. ARBORETUM not only promotes the development of AI models, but also provides critical support for the development of digital tools ranging from pest control to global biodiversity assessment and environmental protection.




