作者= Yadav Ritu Nascetti安德里亚,禁止易纺TITLE =深度的融合网络洪水检测uni-temporal Sentinel-1数据=遥感前沿》杂志上体积= 3 = 2022年URL = //www.thespel.com/articles/10.33雷竞技rebat89/frsen.2022.1060144 DOI = 10.3389 /抽象frsen.2022.1060144 ISSN = 2673 - 6187 =洪水发生在全球范围内,由于气候变化,洪水预计将增加在未来几年。现状敦促更多关注有效监测洪水和检测领域的影响。在这项研究中,我们提出两个分割网络对洪水检测uni-temporal Sentinel-1合成孔径雷达数据。第一个网络是“细心U-Net”。这需要VV, VH, VV / VH作为输入。网络使用空间和channel-wise注意增强特征图,帮助学习更好的分割。“细心U-Net”收益率67%交叉在联盟(借据)Sen1Floods11数据集,3%比基准的借据。第二个提出网络dual-stream“融合网络”,我们融合全球低分辨率的高程数据和永久水面具Sentinel-1 (VV, VH)数据。在Sen1Floods11前面的基准数据集相比,我们的融合网络给了4.5%更好的借据得分。定量,提出方法的性能改进是相当大的。 The quantitative comparison with the benchmark method demonstrates the potential of our proposed flood detection networks. The results are further validated by qualitative analysis, in which we demonstrate that the addition of a low-resolution elevation and a permanent water mask enhances the flood detection results. Through ablation experiments and analysis we also demonstrate the effectiveness of various design choices in proposed networks. Our code is available on Github at https://github.com/RituYadav92/UNI_TEMP_FLOOD_DETECTION以便重用。