Lightweight_Parallel_Deep_Learning_via_Task_Decoupling_for_LargeScale_SAR_Lake_Segmentation

This dataset supports the research presented in the paper “Lightweight Parallel Deep Learning via Task Decoupling for Large-Scale SAR Lake Segmentation”. The study proposes a lightweight parallel deep learning framework designed for large-scale lake extraction from SAR imagery. The dataset includes:(1) lake segmentation results generated by the proposed model,(2) the source code of the proposed method,(3) lake extraction results produced by the U-Net model used in comparative experiments, and(4) lake identification results derived from the NDVI-based method. These materials are provided to support the findings of the study and enable reproducibility and further research in SAR-based lake segmentation.