HR-MTCD（High-Resolution Multi-Temporal Change Detection）

This dataset is mainly composed of high-resolution optical images with sub meter spatial resolution (about 0.5 meters), spanning five years in time. The samples are organized according to the principle of annual annotation and continuous alignment. The dataset contains 1601 sets of images, totaling 8005 images, covering multiple typical regions spatially. The annotation adopts a full element binary change strategy, and this dataset supports both dual temporal change detection and multi temporal modeling and trend analysis, providing an important data foundation for remote sensing change detection from discrete dual temporal change recognition to multi temporal surface change characterization.