In recent years,under the deep integration of artificial intelligence technology and remote sensing big data,change detection frameworks based on deep learning have demonstrated excellent performance through extensive training with annotated data.However,the annotation of change detection data requires pixel-to-pixel comparison of differences between two images,which incurs significant human and time costs.To address the limitations of data annotation,semi-supervised learning-based change detection frameworks have gradually become a research hotspot.This framework can fully leverage a large amount of unlabeled data to enhance the robustness of change detection methods and reduce the model's dependence on annotated data.
维普
万方数据