融合SAR的光学遥感影像双激活门控卷积厚云去除  

Dual-activation gated convolution with SAR fusion for thick cloud removal from optical remote sensing images

作　　者：杨金辉 高贤君[1] 寇媛[2,5] 于盛妍 许磊 杨元维 YANG Jinhui;GAO Xianjun;KOU Yuan;YU Shengyan;XU Lei;YANG Yuanwei(School of Geosciences,Yangtze University,Wuhan 430100,China;Innovation and Research Department,the First Surveying and Mapping Institute of Hunan Province,Changsha 421001,China;Inner Mongolia Autonomous Region Surveying and Mapping Geographic Information Center,Hohhot 010050,China;China Railway Design Corporation,Tianjin 300308,China;Hunan Engineering Research Center of 3D Real Scene Construction and Application Technology,Changsha 421001,China)

机构地区：[1]长江大学地球科学学院,湖北武汉430100 [2]湖南省第一测绘院创新研发部,湖南长沙421001 [3]内蒙古自治区测绘地理信息中心,内蒙古自治区呼和浩特010050 [4]中国铁路设计集团有限公司,天津300308 [5]实景三维建设与应用技术湖南省工程研究中心,湖南长沙421001

出　　处：《浙江大学学报(工学版)》2025年第4期804-813,841,共11页Journal of Zhejiang University:Engineering Science

基　　金：西藏自治区科技计划重大专项(XZ202402ZD0001);深地国家科技重大专项(2024ZD1001003);城市轨道交通数字化建设与测评技术国家工程实验室开放课题基金资助项目(2023ZH01);天津市科技计划项目(23YFYSHZ00190,23YFZCSN00280);湖南省自然科学基金项目部门联合基金资助项目(2024JJ8327).

摘　　要：针对现有基于深度学习的去云方法性能不稳定、色调不均衡等问题,提出融合合成孔径雷达(SAR)和光学数据的遥感影像厚云去除网络.利用SAR图像的真实纹理信息和光学图像的空谱特征信息,从全局和局部构建特征重建任务,引导网络重建云遮挡区域的缺失信息.使用双激活门控卷积块和通道注意力块构建空谱特征推理重建模块,提高网络对非云区域有用信息的特征提取能力.根据云形态和含云量不同将SEN12MS-CR-TS数据集拆分成4个子数据集进行训练和测试.实验结果表明,此方法的峰值信噪比(PSNR)和结构相似性指数(SSIM)比去云效果最优的对比方法分别高出1.0384 dB和0.0915,说明融合SAR和光学数据的遥感影像厚云去除网络可有效去除影像中的云,并完成云下细节信息的重建.A cloud removal network for remote sensing imagery that integrated synthetic aperture radar(SAR)and optical data was proposed to address the issues of unstable performance and uneven color tones in existing deep learning-based cloud removal methods.The true texture information from SAR images and the spatial-spectral feature information from optical images were used to construct feature reconstruction tasks both globally and locally,and these tasks guided the network to rebuild missing information in cloud-covered areas.The dual-activation gated convolutional blocks and the channel attention blocks were utilized to build a spatial-spectral feature inference and reconstruction block which significantly enhanced the network’s ability to extract features from useful information in non-cloud areas.The SEN12MS-CR-TS dataset was divided into four subsets based on different cloud morphologies and cloud contents for training and testing.The experimental results showed that the peak signal-to-noise ratio(PSNR)and structural similarity index measure(SSIM)of the proposed method were 1.0384 dB and 0.0915,respectively,which were higher than those of the best cloud removal methods.Thus the remote sensing image thick cloud removal network,which integrates SAR and optical data,can effectively remove clouds from images and reconstruct the details beneath the clouds.

关 键 词：厚云去除 SAR 数据融合 门控卷积 双流引导 

分 类 号：P237[天文地球—摄影测量与遥感] TP751[天文地球—测绘科学与技术]