JTRNEO/SyntheWorld

SyntheWorld是一个用于土地覆盖制图和建筑物变化检测任务的大规模合成数据集。该数据集在质量、多样性和规模上无与伦比，包含40,000张亚米级像素的图像，具有八类细粒度土地覆盖注释和40,000对双时相图像对，用于建筑物变化检测任务。数据集还通过多个基准遥感数据集上的实验验证了其有效性。文件结构包括不同尺寸和地面采样距离的图像和掩码，分别用于土地覆盖制图和建筑物变化检测任务。

SyntheWorld is a large-scale synthetic dataset designed for land cover mapping and building change detection tasks. The dataset is unparalleled in quality, diversity, and scale, featuring 40,000 images with submeter-level pixels, fine-grained land cover annotations of eight categories, and 40,000 pairs of bitemporal image pairs for building change detection tasks. The datasets effectiveness has been validated through experiments on multiple benchmark remote sensing datasets. The file structure includes images and masks of different sizes and ground sampling distances (GSD) for land cover mapping and building change detection tasks.