|本期目录/Table of Contents|

[1]汪家明,卢 涛*.多尺度残差深度神经网络的卫星图像超分辨率算法[J].武汉工程大学学报,2018,40(04):440-445.[doi:10. 3969/j. issn. 1674?2869. 2018. 04. 018]
 WANG Jiaming,LU Tao *. Satellite Imagery Super-Resolution Algorithm via Multi-Scale Residual Deep Neural Network[J].Journal of Wuhan Institute of Technology,2018,40(04):440-445.[doi:10. 3969/j. issn. 1674?2869. 2018. 04. 018]
点击复制

多尺度残差深度神经网络的卫星图像超分辨率算法(/HTML)
分享到:

《武汉工程大学学报》[ISSN:1674-2869/CN:42-1779/TQ]

卷:
40
期数:
2018年04期
页码:
440-445
栏目:
机电与信息工程
出版日期:
2018-08-23

文章信息/Info

Title:

Satellite Imagery Super-Resolution Algorithm via Multi-Scale Residual Deep Neural Network
文章编号:
20180418
作者:
汪家明卢 涛*

武汉工程大学计算机科学与工程学院,湖北 武汉 430205
Author(s):
WANG Jiaming LU Tao *

School of Computer Science and Engineering,Wuhan Institute of Technology,Wuhan 430205, China
关键词:
卫星图像超分辨率残差网络残差学习卷积神经网络
Keywords:
satellite imagesuper-resolutionresidual networkmulti-scale imageconvolutional neural network
分类号:
TP391.4
DOI:
10. 3969/j. issn. 1674?2869. 2018. 04. 018
文献标志码:
A
摘要:

卫星图像实现星际对地观测并被广泛的应用到了军事和经济生活领域。受到星载成像设备和星地通讯带宽的限制,卫星图像的地面分辨率常不能完全满足目标识别与分析的需求。卫星图像的成像幅度宽且范围广,地面目标的尺度变化大、纹理信息多样化,给现有图像超分辨率技术带来了新的挑战。针对卫星图像的多尺度特性,提出了一种多尺度残差深度神经网络,首先提取低分辨率卫星图像的多尺度特征,对不同尺度特征建立自适应深度神经网络,然后使用融合网络进行残差融合,融合不同尺度高频信息,最终生成高分辨卫星图像。在SpaceNet卫星图像数据集中的实验结果证明了本文算法的优越性。
Abstract:

Satellite imagery realizes interstellar-earth observations, which is widely used in military and economic fields. Because the performances of satellite-borne imaging equipment and the band width of satellite communications system are limited, the resolution of ground targets in satellite images are often low, thus they cannot fully meet the needs of target identification and analysis. Moreover, satellite images have three features: wide range of imaging, variation of multi-scale of ground targets, and diversification of texture information, which bring new challenges to the existing super-resolution algorithms. Using the multi-scale nature of satellite image, a multi-scale residual neural network was proposed in this paper for accurately reconstructing the multi-scale information. Firstly, different scale features of low-resolution satellite images were extracted, then for each scale-level, an adaptive deep residual neural network was developed for better reconstruction performance. Then a fusion network was used to refine different scales of residual information. The proposed fusion network fuses high-frequency information of different scales to output the target high-resolution satellite image. Experimental results over the SpaceNet satellite image database prove the superiority of the proposed algorithm.

参考文献/References:


[1] 杨欣, 费树岷, 周大可. 基于MAP的自适应图像配准及超分辨率重建[J]. 仪器仪表学报, 2011, 32(8):1771-1775. [2] 夏平平, 吕太之. 动态人脸识别系统的设计与实现[J]. 武汉工程大学学报, 2011, 33(10):107-110. [3] 赵娜, 赵彤洲, 邹冲,等. 稀疏表示中字典学习的影响因子研究[J]. 武汉工程大学学报, 2017, 39(3):267-272. [4] 肖杰, 卢涛, 万永静,等. 基于稀疏表达的图像超分辨率算法实现[J]. 计算机与数字工程, 2014(11):2160-2163. [5] 王宏, 卢芳芳, 李建武. 结合支持向量回归和图像自相似的单幅图像超分辨率算法[J]. 中国图象图形学报, 2016, 21(8):986-992. [6] 吴炜, 杨晓敏, 陈默,等. 基于偏最小二乘算法的人脸图像超分辨率技术[J]. 光子学报, 2009, 38(11):3025-3033. [7] DONG C, LOY C C, HE K, et al. Image super- resolution using deep convolutional networks[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2016, 38(2): 295-307. [8] LU T,WANG H,XIONG Z,et al. Face hallucination using region-based deep convolutional networks[C]// IEEE International Conference on Image Processing. Beijing: IEEE,2017:1657-1661.[9] SHI W,CABALLERO J,HUSZ?R F,et al. Real-time single image and video super-resolution using an efficient sub-pixel convolutional neural network [C]// Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas: IEEE,2016:1874-1883.[10] KIM J, KWON L J, MU L. Accurate image super- resolution using very deep convolutional networks[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas: IEEE,2016: 1646-1654.[11] LEDIG C, THEIS L, HUSZ?R F, et al. Photo- realistic single image super-resolution using a generative adversarial network[J]. Computer Vision and Pattern Recognition, 2017(2):105-114.[12] LUO Y, ZHOU L, WANG S, et al. Video satellite imagery super resolution via convolutional neural networks[J]. IEEE Geoscience & Remote Sensing Letters, 2017, 14(12):2398-2402. [13] 宋晓宇, 单新建. 高分辨率卫星影像在城市建筑物识别中的初步应用[J]. 遥感信息, 2002(1):27-31. [14] HE K, ZHANG X, REN S, et al. Deep Residual Learning for Image Recognition[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas: IEEE,2016:770-778.[15] 卢涛, 章瑾, 陈白帆,等. 多尺度自适应配准的视频超分辨率算法[J]. 武汉工程大学学报, 2016, 38(2):178-184. [16] DONG C, LOY C C, TANG X. Accelerating the super-resolution convolutional neural network[C]//European Conference on Computer Vision. Amsterder:Springer, 2016: 391-407. [17] LAI W S, HUANG J B, AHUJA N, et al. Deep laplacian pyramid networks for fast and accurate super-resolution[C]//Proc. IEEE Conf. Comput. Vis. Pattern Recognition. 2017: 624-632.

相似文献/References:

备注/Memo

备注/Memo:

收稿日期:2018-04-16基金项目:国家自然科学基金(61502354,61671332,41501505);湖北省自然科学基金(2015CFB451,2014CFA130,2012FFA099,2012FFA134,2013CF125);武汉工程大学科研基金(K201713)作者简介:汪家明,硕士研究生。E-mail:549880890@qq.com *通讯作者:卢 涛,博士,副教授。E-mail:lut@wit.edu.cn引文格式:汪家明,卢涛. 基于多尺度残差深度神经网络的卫星图像超分辨率算法[J]. 武汉工程大学学报,2018,40(4):440-445.
更新日期/Last Update: 2018-08-16