首页 > 论文 > 光学 精密工程 > 22卷 > 7期(pp:1886-1895)

联合块匹配与稀疏表示的卫星云图修复

Satellite cloud image inpainting based on patch matching and sparse representation

  • 摘要
  • 论文信息
  • 参考文献
  • 被引情况
  • PDF全文
分享:

摘要

针对卫星云图在接收及传输过程中受噪声、大气湍流、太阳风暴及卫星轨道漂移等影响造成的云图数据破损, 提出了一种联合块匹配与稀疏表示的卫星云图修复方法。首先, 根据破损区域的优先权值确定待修复像素, 对该像素的邻域进行分块处理。然后, 利用待修复块与各匹配块之间的结构相似度, 建立相应的冗余字典; 通过求解稀疏表示问题修复该破损区域。最后, 沿着等照度线不断更新优先权值, 实现整幅图像的修复。实验结果表明, 提出的方法不仅能避免传统偏微分方程(PDE)修复法所导致的结构丢失, 也能很好地改善基于纹理填充修复方法所导致的修复不足及块效应现象。测试结果显示: 在云图存在局部区域缺失时, 修复后云图的峰值信噪比(PSNR)比匹配追踪法及总变分法的修复结果平均提高了8.50 dB和0.28 dB, 而且在纹理细节及边缘区域具有更好的视觉效果。

Abstract

For some defects of satellite cloud images caused by noises, atmospheric turbulence, solar storms and satellite orbit drifts in the receiving and transmission process, a novel satellite cloud image inpainting method using patch matching and sparse representation was proposed. Firstly, a pixel to be inpainted was searched according to the priority of the damaged area. The neighborhood of this pixel was divided into patches, and a redundant dictionary was constructed by calculating the structural similarity of inpainting patch and matching patches. Then the damaged area was repaired by solving a sparse representation problem. Finally, the whole cloud image was inpainted by updating priority along an isophote repeatedly. The experimental results show that the proposed method can not only avoid the structure missing from the traditional Partial Differential Equation ( PDE) method, but also can improve the texture details and blocking effect of the texture filling repair method. For the cloud image with local defects, the inpainted image by proposed method can improve Peak Signal to Noise Ratio(PSNR) by 8.50 dB and 0.28 dB as compared to Matching Pursuit (MP) method and Total Variation (TV) method respectively. It shows better visual effects on texture details and edge regions.

Newport宣传-MKS新实验室计划
补充资料

中图分类号:TP75

DOI:10.3788/ope.20142207.1886

所属栏目:信息科学

基金项目:国家自然科学基金资助项目(No.61271399, No.61373068); 浙江省自然科学基金资助项目(No.Y1111061); 宁波市自然科学基金资助项目 (No.2011A610192,No.2013A610055); 宁波市科技创新团队研究计划资助项目(No.2011B81002); 宁波大学研究生教育改革研究重点项目(No.JGZDI201202);宁波大学科研基金资助项目(No.XYL12003,No.XKXL1306)

收稿日期:2013-10-10

修改稿日期:2013-11-29

网络出版日期:--

作者单位    点击查看

金炜:宁波大学 信息科学与工程学院, 浙江 宁波 315211
王文龙:宁波大学 信息科学与工程学院, 浙江 宁波 315211
符冉迪:宁波大学 信息科学与工程学院, 浙江 宁波 315211
田文哲:宁波大学 信息科学与工程学院, 浙江 宁波 315211
尹曹谦:宁波大学 信息科学与工程学院, 浙江 宁波 315211

联系人作者:金炜(xyjw1969@126.com)

备注:金炜(1969-), 男, 浙江兰溪人, 博士, 副教授, 硕士生导师, 2006年于重庆大学获得博士学位, 主要从事遥感图像处理、多尺度几何分析、压缩感知、光电检测的研究。

【1】赵姼, 周其永, 毛成忠.FY卫星系统云图失真问题的一种解决方法[J].气象水文海洋仪器, 2011, 2(6):114-118.
ZHAO SH, ZHOU Q Y, MAO CH ZH.A method for solving abnormal images in FY meteorological satellite system[J].Meteorological, Hydrological and Marine Instruments, 2011, 2(6):114-118. (in Chinese)

【2】朱长明, 沈占锋, 骆剑承, 等.基于MODIS数据的Landsat-7 SLC-off影像修复方法研究[J].测绘学报, 2010, 39(6):251-256.
ZHU CH M, SHEN ZH F, LOU J CH, et al..Research on landsat-7 SLC-off image restoration method based on MODIS09 data [J].Acta Geodaetica et Cartographica Sinica, 2010, 39(6):251-256. (in Chinese)

【3】吕恒毅, 刘杨, 薛旭成.遥感图像星上背景扣除和灰度拉伸方案与实验[J].液晶与显示, 2012, 27(2):235-240.
LV H Y, LIU Y, XUE X CH.Methods and experiments of background subtraction and grayscale stretch for remote sensing images [J].Chinese Journal of Liquid Crystals and Displays, 2012, 27(2):235-240. (in Chinese)

【4】BERTALMIO M, SAPIRO G, CASELLES V, et al.. Proceedings of international conference on computer graphics and interactive techniques[C]. New Orleans, Louisiana USA, 2000, 1:417-424.

【5】CRIMINISI A, PEREA P, TOYAMA K. Region filling and object removal by exemplar-based image inpainting[J]. IEEE Transactions on Image Processing, 2004, 13(9): 1200-1212.

【6】冯亮, 王平, 许廷发, 等.运动模糊退化图像的双字典稀疏复原[J].光学精密工程, 2011, 19(8):1982-1989.
FENG L, WANG P, XU T F, et al.. Dual dictionary sparse restoration of blurred images[J].Opt. Precision Eng., 2011, 19(8):1982-1989. (in Chinese)

【7】ZHOU M, CHEN H, PAISLEY J, et al.. Nonparametric Bayesian dictionary learning for analysis of noisy and incomplete images[J]. IEEE Transactions on Image Processing, 2012, 21(1):130-144.

【8】曾文静, 万磊, 张铁栋, 等.复杂海空背景下弱小目标的自动检测[J].光学精密工程, 2012, 20(2):403-413.
ZENG W J, WAN L, ZHANG T D, et al.. Fast detection of weak targets in complex sea-sky background [J]. Opt. Precision Eng., 2012, 20(2):403-413. (in Chinese)

【9】李民, 程建, 李小文, 等.非局部学习字典的图像修复[J].电子与信息学报, 2011, 33(11): 2672-2678.
LI M, CHENG J, LI X W, et al..Image inpainting based on non-local learned dictionary[J]. Journal of Electronics & Information Technology, 2011, 33(11): 2672-2678. (in Chinese)

【10】李长洋.基于稀疏性的图像分层修复[D].成都: 西南交通大学, 2010.
LI CH Y.Image inpainting based on decomposition[D].Chengdu:XiNan jiaotong University, 2010. (in Chinese)

【11】刘洋, 田小建, 王晴, 等.采用局部分形的高效图像分割方法在红外云图处理中的应用[J].光学精密工程, 2011, 19(6):1367-1375.
LIU Y, TIAN X J, WANG Q, et al.. Application of efficient image segmentation method based on local fractal in the infrared cloud image process [J].Opt. Precision Eng., 2011, 19(6):1367-1375. (in Chinese)

【12】WANG ZH, BOVIK A C, SHEIKH H R, et al.. Image quality assessment from error visibility to structural similarity [J]. IEEE Transactions on Image Processing, 2004, 13(4): 600-612.

【13】AHARON M, ELAD M, BRUCKSTEIN A. K-SVD:An algorithm for designing overcomplete dictionaries for sparse representation[J].IEEE Trans.Signal Process, 2006, 54(11):4311-4322.

【14】吴君钦, 李艳丽, 刘昊.“类整数DCT”变换基去相关性能分析[J].液晶与显示, 2013, 28(2):278-284.
WU J Q, LI Y L,LIU H.De-correlation characteristic analysis of variety integer DCT transform radix [J].Chinese Journal of Liquid Crystals and Displays, 2013, 28(2):278-284. (in Chinese)

【15】GETREUER P. RED: tvreg v2: Variational Imaging Methods for Denoising, Deconvolution, Inpainting, and Segmentation[OL].http://www.mathworks.com/matlabcentral/fileexchange/29743-tvreg.2012.11.

您的浏览器不支持PDF插件,请使用最新的(Chrome/Fire Fox等)浏览器.或者您还可以点击此处下载该论文PDF