MLRS Net

Description: MLRSNet is a multi-label high spatial resolution remote sensing dataset designed for semantic scene understanding. It offers diverse perspectives of the world captured from satellites, comprising high spatial resolution optical satellite images. The dataset contains 109,161 remote sensing images, meticulously annotated into 46 categories, with each category holding between 1,500 to 3,000 sample images. Each image maintains a fixed size of 256×256 pixels, featuring various pixel resolutions ranging from approximately 10 meters to 0.1 meters. Download Dataset Each image in MLRSNet is tagged with several of 60 predefined class labels, with the number of labels per image varying between 1 and 13. This comprehensive labeling enables the dataset to support multiple applications, including multi-label based image classification, multi-label based image retrieval, and image segmentation. MLRSNet thus serves as a valuable resource for advancing research and development in remote sensing and related fields. This dataset is sourced from Kaggle.数据集说明： MLRSNet是一款专为语义场景理解设计的多标签高空间分辨率遥感数据集，收录了卫星采集的多样化全球视角影像，全部为高空间分辨率光学卫星影像。该数据集包含109161张遥感影像，经精细标注划分为46个类别，每个类别涵盖1500至3000张样本影像。所有影像均采用256×256像素的固定尺寸，像素分辨率范围约为10米至0.1米。 数据集下载 MLRSNet中的每张影像均被标注了60个预定义类别标签中的若干个，单张影像的标签数量在1至13之间浮动。这种全面的标注方式使得该数据集可支撑多种应用场景，包括基于多标签的图像分类、基于多标签的图像检索以及图像分割任务。因此，MLRSNet可作为推动遥感及相关领域研究与开发的宝贵资源。 本数据集源自Kaggle平台。