用于遥感图像超分辨率重建的残差对偶回归网络

doi:10.6046/zrzyyg.2021208

摘要
图/表
参考文献
相关文章
Metrics

全文: PDF(3859 KB)

HTML
输出: BibTeX | EndNote (RIS)

摘要

使用人工模拟的高-低分辨率图像对易导致在对真实遥感图像超分辨率重建时模型泛化能力差,针对此问题,结合残差通道注意力网络(residual channel attention network,RCAN)的二次残差(residual in residual,RIR)模块,改进对偶回归网络(dual regression networks,DRN),提出了残差对偶回归网络(residual dual regression network,RDRN)。选取LandCover.ai和DIOR航空图像数据集的10 000张512像素×512像素图像构成样本数据集,用于训练和测试网络,并将重建结果与现有其他超分辨率网络模型的重建结果对比评价。实验结果表明,RDRN在重建质量和模型参数量方面均表现优异,能够在较低模型复杂度的情况下实现较好的超分重建效果,且对不同低分辨率遥感图像具有较好的泛化能力。

	服务

	把本文推荐给朋友
	加入引用管理器
	E-mail Alert
	RSS
	作者相关文章
	尚晓梅
	李佳田
	吕少云
	杨汝春
	杨超

关键词 ：遥感图像, 超分辨率重建, 对偶回归, 残差通道注意力, 闭合网络

Abstract：

In order to solve the problem of poor model generalizing ability in real super-resolution reconstruction of remote sensing images, which is easily caused by the use of artificial high-low resolution image pairs, combined with the residual in residual (RIR) module of residual channel attention network (RCAN), dual regression network (DRN) is improved, and residual dual regression network (RDRN) is proposed. Ten thousand 512 × 512 pixel images from LandCover.ai and DIOR aerial image data sets were selected to form the sample data set for training and testing the network, and the reconstruction results were compared with those of other super-resolution network models. The experimental results show that RDRN has an excellent performance in both reconstruction quality and model parameters. It can achieve a better super segmentation reconstruction effect with lower model complexity and has good generalization ability for different low-resolution remote sensing images.

Key words： remote sensing image super-resolution reconstruction dual regression residual channel attention closed network

收稿日期: 2021-06-30 出版日期: 2022-06-20

ZTFLH:

TP79

基金资助:国家自然科学基金项目“城市居住区识别的Voronoi邻域方法与初步实践”(41561082)

通讯作者: 李佳田

作者简介: 尚晓梅(1997-),女,硕士研究生,主要研究方向为摄影测量与遥感。Email: sxm.320@qq.com。

引用本文:

尚晓梅, 李佳田, 吕少云, 杨汝春, 杨超. 用于遥感图像超分辨率重建的残差对偶回归网络[J]. 自然资源遥感, 2022, 34(2): 112-120.
SHANG Xiaomei, LI Jiatian, LYU Shaoyun, YANG Ruchun, YANG Chao. Residual dual regression network for super-resolution reconstruction of remote sensing images. Remote Sensing for Natural Resources, 2022, 34(2): 112-120.

链接本文:

https://www.gtzyyg.com/CN/10.6046/zrzyyg.2021208 或 https://www.gtzyyg.com/CN/Y2022/V34/I2/112

Fig.1 对偶学习SR重建模型

Fig.2 二次残差块的结构

Fig.3 RDRN的网络结构

Fig.4 镜像加边操作

Fig.5 不同尺度SR重建的视觉比较

Fig.6 不同SR方法的视觉比较

Tab.1 不同SR算法的性能比较^①

Tab.2 模型参数量对比

Tab.3 对偶残差方法对SR质量的影响

Tab.4 对偶损失权重λ对SR性能的影响

Fig.8 不成对数据占比p对4倍重建SR性能的影响

[1]	张艳, 卢宣铭, 刘国瑞, 等. 多路径特征融合的遥感图像超分辨率重建算法[J]. 遥感信息, 2021, 36(2):46-53.
	Zhang Y, Lu X M, Liu G R, et al. Super resolution reconstruction algorithm of remote sensing image based on multi-path feature fusion[J]. Remote Sensing Information, 2021, 36(2):46-53.
[2]	Harris J L. Diffraction and resolving power[J]. Journal of the Optical Society of America, 1964, 54(7):931-936. doi: 10.1364/JOSA.54.000931
[3]	苏衡, 周杰, 张志浩. 超分辨率图像重建方法综述[J]. 自动化学报, 2013, 39(8):1202-1213.
	Su H, Zhou J, Zhang Z H. Review of super resolution image reconstruction methods[J]. Acta Automatica Sinica, 2013, 39(8):1202-1213. doi: 10.3724/SP.J.1004.2013.01202
[4]	Dong C, Loy C C, He K M, et al. Learning a deep convolutional network for image super-resolution[C]// David Fleet.Proceedings of European Conference on Computer Vision. Zurich: Computer Vision-ECCV, 2014,184-199.
[5]	Kim J, Lee J K, Lee K M. Accurate image super-resolution using very deep convolutional networks[C]// IEEE.Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. New York: IEEE, 2016,1646-1654.
[6]	He K M, Zhang X Y, Ren S Q, et al. Deep residual learning for image recognition[C]// IEEE.Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas: IEEE, 2016,770-778.
[7]	Kim J, Lee J K, Lee K M. Deeply-recursive convolutional network for image super-resolution[C]// IEEE.Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas: IEEE, 2016:1637-1645.
[8]	Tai Y, Yang J, Liu X. Image super-resolution via deep recursive residual network[C]// IEEE.Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.Honolulu:IEEE, 2017:3147-3155.
[9]	刘帅, 朱亚杰, 薛磊. 一种结合稀疏表示和纹理分块的遥感影像超分辨率方法[J]. 武汉大学学报(信息科学版), 2015, 40(5):578-582.
	Liu S, Zhu Y J, Xue L. A super resolution method for remote sensing image based on sparse representation and texture segmentation[J]. Geomatics and Information Science of Wuhan University, 2015, 40(5):578-582.
[10]	吴琼, 田越, 周春平, 等. 遥感图像超分辨率研究的现状和发展[J]. 测绘科学, 2008, 33(6):66-69,15.
	Wu Q, Tian Y, Zhou C P, et al. Current situation and development of research on super resolution of remote sensing image[J]. Surveying and Mapping Science, 2008, 33(6):66-69,15.
[11]	Shocher A, Cohen N, Irani M. “Zero-shot” super-resolution using deep internal learning[C]// IEEE.Proceedings of the 2018 IEEE Conference on Computer Vision and Pattern Recognition, Salt Lake City: IEEE, 2018:3118-3126.
[12]	Bulat A, Yang J, Tzimiropoulos G. To learn image super-resolution,use a GAN to learn how to do image degradation first[C]// Vittorio Ferrari.Proceedings of the European Conference on Computer Vision).Munich:ECCV, 2018:185-200.
[13]	Yi Z L, Zhang H, Tan P, et al. DualGAN:unsupervised dual learning for image-to-image translation[C]// IEEE.Proceedings of the IEEE International Conference on Computer Vision.Venice:ICCV, 2017:2868-2876.
[14]	Ledig C, Theis L, Huszár F, et al. Photo-realistic single image super-resolution using a generative adversarial network[C]// IEEE.Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.Honolulu:IEEE, 2017:4681-4690.
[15]	Simonyan K, Zisserman A. Very deep convolutional networks for large-scale image recognition[J/OL]. https://arxiv.org/abS/1409.1556v3.
[16]	Yuan Y, Liu S, Zhang J, et al. Unsupervised image super-resolution using cycle-in-cycle generative adversarial networks[C]// IEEE. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops: IEEE, 2018:701-710.
[17]	Zhu J Y, Park T, Isola P, et al. Unpaired image-to-image translation using cycle-consistent adversarial networks[C]// IEEE.Proceedings of the IEEE International Conference on Computer Vision.Venice:IEEE, 2017:2223-2232.
[18]	陈赛健, 朱远平. 基于生成对抗网络的文本图像联合超分辨率与去模糊方法[J]. 计算机应用, 2020, 40(3):859-864.
	Chen S J, Zhu Y P. Text image joint super resolution and deblurring method based on generative countermeasure network[J]. Computer Applications, 2020, 40(3):859-864.
[19]	李欣, 韦宏卫, 张洪群. 结合深度学习的单幅遥感图像超分辨率重建[J]. 中国图象图形学报, 2018, 23(2):209-218.
	Li X, Wei H W, Zhang H Q. Super resolution reconstruction of single remote sensing image combined with deep learning[J]. Chinese Journal of Image Graphics, 2018, 23(2):209-218.
[20]	Guo Y, Chen J, Wang J, et al. Closed-loop matters:Dual regression networks for single image super-resolution[C]// IEEE.IEEE Conference on Computer Vision and Pattern Recognition.Seattle:CVPR, 2020:5407-5416.
[21]	Zhang Y, Li K, Li K, et al. Image super-resolution using very deep residual channel attention networks[C]// Vittorio Ferrari.Proceedings of the European Conference on Computer Vision.Munich:ECCV, 2018:286-301.
[22]	Fu J, Liu J, Tian H, et al. Dual attention network for scene segmentation[C]// Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Long Beach: IEEE, 2019:3141-3149.
[23]	Yu Z, Qiang Y. An overview of multi-task learning[J]. National Science Review, 2018, 5(1):30-43. doi: 10.1093/nsr/nwx105
[24]	周飞燕, 金林鹏, 董军. 卷积神经网络研究综述[J]. 计算机学报, 2017, 40(6):1229-1251.
	Zhou F Y, Jin L P, Dong J. Review of convolutional neural networks[J]. Chinese Journal of Computers, 2017, 40(6):1229-1251.
[25]	张顺, 龚怡宏, 王进军. 深度卷积神经网络的发展及其在计算机视觉领域的应用[J]. 计算机学报, 2019, 42(3):453-482.
	Zhang S, Gong Y H, Wang J J. Development of deep convolutional neural network and its application in computer vision[J]. Chinese Journal of Computers, 2019, 42(3):453-482.
[26]	Boguszewski A, Batorski D, Ziemba-Jankowska N, et al. LandCover. ai:Dataset for automatic mapping of buildings,woodlands,water and roads from aerial imagery[C]// Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.Nashville:IEEE, 2021:1102-1110.
[27]	Li K, Wan G, Cheng G, et al. Object detection in optical remote sensing images:A survey and a new benchmark[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2020, 159:296-307. doi: 10.1016/j.isprsjprs.2019.11.023
[28]	Wan M, Yazid, et al. A review of image quality assessment (IQA):SNR,GCF,AD,NAE,PSNR,ME[J]. Journal of Advanced Research in Computing and Applications. 2017, 7(1):1-7.

[1]	王建强, 邹朝晖, 刘荣波, 刘志松. 基于U²-Net深度学习模型的沿海水产养殖塘遥感信息提取[J]. 自然资源遥感, 2023, 35(3): 17-24.
[2]	徐欣钰, 李小军, 赵鹤婷, 盖钧飞. NSCT和PCNN相结合的遥感图像全色锐化算法[J]. 自然资源遥感, 2023, 35(3): 64-70.
[3]	牛祥华, 黄微, 黄睿, 蒋斯立. 基于注意力特征融合的高保真遥感图像薄云去除[J]. 自然资源遥感, 2023, 35(3): 116-123.
[4]	胡建文, 汪泽平, 胡佩. 基于深度学习的空谱遥感图像融合综述[J]. 自然资源遥感, 2023, 35(1): 1-14.
[5]	马晓剑, 赵法舜, 刘艳宾. 多特征准则融合的遥感图像脉冲噪声的识别处理[J]. 自然资源遥感, 2022, 34(3): 17-26.
[6]	张大明, 张学勇, 李璐, 刘华勇. 一种超像素上Parzen窗密度估计的遥感图像分割方法[J]. 自然资源遥感, 2022, 34(1): 53-60.
[7]	李轶鲲, 杨洋, 杨树文, 王子浩. 耦合模糊C均值聚类和贝叶斯网络的遥感影像后验概率空间变化向量分析[J]. 自然资源遥感, 2021, 33(4): 82-88.
[8]	刘万军, 高健康, 曲海成, 姜文涛. 多尺度特征增强的遥感图像舰船目标检测[J]. 自然资源遥感, 2021, 33(3): 97-106.
[9]	王小兵. 融合提升小波阈值与多方向边缘检测的矿区遥感图像去噪[J]. 国土资源遥感, 2020, 32(4): 46-52.
[10]	刘尚旺, 崔智勇, 李道义. 基于Unet网络多任务学习的遥感图像建筑地物语义分割[J]. 国土资源遥感, 2020, 32(4): 74-83.
[11]	李宇, 肖春姣, 张洪群, 李湘眷, 陈俊. 深度卷积融合条件随机场的遥感图像语义分割[J]. 国土资源遥感, 2020, 32(3): 15-22.
[12]	蔡之灵, 翁谦, 叶少珍, 简彩仁. 基于Inception-V3模型的高分遥感影像场景分类[J]. 国土资源遥感, 2020, 32(3): 80-89.
[13]	叶发茂, 罗威, 苏燕飞, 赵旭青, 肖慧, 闵卫东. 卷积神经网络特征在遥感图像配准中的应用[J]. 国土资源遥感, 2019, 31(2): 32-37.
[14]	谢奇芳, 姚国清, 张猛. 基于Faster R-CNN的高分辨率图像目标检测技术[J]. 国土资源遥感, 2019, 31(2): 38-43.
[15]	葛芸, 江顺亮, 叶发茂, 姜昌龙, 陈英, 唐祎玲. 聚合CNN特征的遥感图像检索[J]. 国土资源遥感, 2019, 31(1): 49-57.

Viewed

Full text

Abstract

Cited

Shared

Discussed