基于SAR遥感影像的水体提取研究

Research on Water Body Extraction Based on SAR Remote-Sensing Images

使用基于Swin Transformer的深度学习模型,成功应用于Sentinel-1合成孔径雷达(Synthetic Aperture Radar,SAR)遥感图像水体提取任务。通过MSTAR数据集进行预训练和迁移学习,模型在遵义地区的实验中表现出显著的准确性和效率。采用图像预处理和Swin Transformer的窗口注意力机制,成功提高了水体提取性能。实验结果显示,该模型在提取水体矢量信息方面取得了显著成果,为遥感图像处理和水体提取领域的进一步研究提供了有力支持。

This study successfully applies a deep learning model based on Swin Transformer to the water body extraction task of remote-sensing Sentinel-1 Synthetic Aperture Radar(SAR) images.The MSTAR dataset is used for pre-training and transfer learning,and the model demonstrates notable accuracy and efficiency in experiments in the Zunyi region.The image preprocessing and the windowed attention mechanism of Swin Transformer are used to successfully improve the performance of water body extraction.Experimental results indicate that the model achieves remarkable results in extracting the vector information of the water body,which provides strong support for further research in the field of remote-sensing image processing and water body extraction.