MM-LUCAS
收藏资源简介:
MM-LUCAS是一个专为评估多模态大语言模型(MM-LLMs)在农业应用中的表现而设计的多模态农业数据集。该数据集基于欧洲联盟(EU)的Land Use/Cover Area Frame Survey (LUCAS)数据集,包含1784张景观图像、分割掩码、深度图以及详细的注释信息。数据集的创建旨在解决农业知识型数据集的局限性问题,并推动农业领域MM-LLMs的发展。MM-LUCAS数据集的内容丰富,涵盖了地理位置、国家、日期、土地覆盖和土地使用分类等详细信息,适用于农业领域的多种复杂任务,如作物分类、病害检测和环境影响预测等。
MM-LUCAS is a multimodal agricultural dataset specifically developed to evaluate the performance of multimodal large language models (MM-LLMs) in agricultural applications. This dataset is built upon the Land Use/Cover Area Frame Survey (LUCAS) dataset of the European Union (EU), and contains 1,784 landscape images, segmentation masks, depth maps, as well as detailed annotation information. The creation of the MM-LUCAS dataset aims to address the limitations of existing knowledge-intensive agricultural datasets and promote the development of MM-LLMs in the agricultural field. The MM-LUCAS dataset has rich content, covering detailed information such as geographic location, country, collection date, land cover and land use categories, and is applicable to a variety of complex agricultural tasks including crop classification, disease detection, and environmental impact prediction.

- 1AgriBench: A Hierarchical Agriculture Benchmark for Multimodal Large Language Models莱布尼茨农业景观研究中心 · 2024年



