LandDiscover50K
收藏资源简介:
LandDiscover50K是由大连理工大学开发的一个综合性遥感图像语义分割数据集,旨在解决开放词汇遥感图像语义分割(OVRSISS)任务中的数据缺乏问题。该数据集包含51,846张高分辨率遥感图像,涵盖了40个不同的语义类别,数据来源包括多个公开的遥感数据集。LandDiscover50K通过整合多样化的传感器、分辨率和地物类别,增强了模型的鲁棒性和泛化能力。该数据集的应用领域主要集中在遥感图像分析,旨在提升农业产量、减轻自然灾害和管理土地覆盖变化等方面的应用效果。
LandDiscover50K is a comprehensive remote sensing image semantic segmentation dataset developed by Dalian University of Technology, which aims to address the data scarcity issue in the task of open-vocabulary remote sensing image semantic segmentation (OVRSISS). This dataset contains 51,846 high-resolution remote sensing images covering 40 distinct semantic categories, with data sourced from multiple public remote sensing datasets. By integrating diverse sensors, spatial resolutions and ground object categories, LandDiscover50K enhances the robustness and generalization performance of models. The dataset is primarily applied in remote sensing image analysis, with the goal of improving the performance of applications such as boosting agricultural production, mitigating natural disasters and managing land cover changes.




