基于卷积神经网络的高分六号卫星多光谱图像压缩

doi:10.19678/j.issn.1000-3428.0064845

计算机工程 ›› 2023, Vol. 49 ›› Issue (9): 287-294. doi: 10.19678/j.issn.1000-3428.0064845

基于卷积神经网络的高分六号卫星多光谱图像压缩

朱孟栩¹^,², 张文豪¹^,²^,*, 李国洪¹^,², 顾行发³, 余涛³^,⁴, 郑逢杰⁵, 张丽丽³^,⁴, 吴俣⁶, 邴芳飞¹^,², 唐健雄¹^,²

1. 北华航天工业学院遥感信息工程学院, 河北廊坊 065000
2. 河北省航天遥感信息处理与应用协同创新中心, 河北廊坊 065000
3. 中国科学院空天信息创新研究院遥感卫星应用国家工程实验室, 北京 100094
4. 中科空间信息(廊坊)研究院, 河北廊坊 065001
5. 航天工程大学航天信息学院, 北京 101416
6. 天津大学地球系统科学学院, 天津 300072

收稿日期:2022-05-30 出版日期:2023-09-15 发布日期:2022-11-14
通讯作者: 张文豪
作者简介:
朱孟栩（1996—），男，硕士研究生，主研方向为图像压缩、定量遥感
李国洪，教授、博士
顾行发，研究员、博士
余涛，研究员、博士
郑逢杰，副教授、博士
张丽丽，助理研究员、博士
吴俣，副教授、博士
邴芳飞，硕士研究生
唐健雄，硕士研究生
基金资助:
国家自然科学基金(41907192); 河北省自然科学基金(D2020409003); 河北省高等学校科学技术研究项目(ZD2021303); 北华航天工业学院博士科研启动基金(BKY-2021-31); 高分辨率对地观测系统重大专项(30-Y30F06-9003-20/22); 国家重点研发计划(2019YFE0127300); 国家重点研发计划(2019YFE0126600); 民用航天预研项目(D040102); 国防基础科研项目(JCKY2020908B001)

GF-6 Multispectral Image Compression Based on Convolutional Neural Network

Mengxu ZHU¹^,², Wenhao ZHANG¹^,²^,*, Guohong LI¹^,², Xingfa GU³, Tao YU³^,⁴, Fengjie ZHENG⁵, Lili ZHANG³^,⁴, Yu WU⁶, Fangfei BING¹^,², Jianxiong TANG¹^,²

1. School of Remote Sensing and Information Engineering, North China Institute of Aerospace Engineering, Langfang 065000, Hebei, China
2. Heibei Spacer Remote Sensing Information Processing and Application of Collaborative Innovation Center, Langfang 065000, Hebei, China
3. National Engineering Laboratory of Remote Sensing Satellite Application, Institute of Aerospace Information Innovation, Chinese Academy of Sciences, Beijing 100094, China
4. China Academy of Spatial Information (Langfang), Langfang 065001, Hebei, China
5. School of Aerospace Information, Aerospace Engineering University, Beijing 101416, China
6. School of Earth System Science, Tianjin University, Tianjin 300072, China

Received:2022-05-30 Online:2023-09-15 Published:2022-11-14
Contact: Wenhao ZHANG

摘要/Abstract

摘要：

高分六号多光谱图像的空间冗余和谱间冗余较高，为了降低高分六号多光谱图像所占用的存储空间，提高国产高分多光谱图像的压缩效率，提出一种基于卷积神经网络的端到端多光谱图像压缩模型SMIC。SMIC模型由自编码器、量化结构、熵编码3个部分组成。自编码器通过卷积下采样提取图像的特征，降低数据的空间维度。量化结构采用多进制量化将特征矩阵离散化，减少图像压缩过程中的信息损失。熵编码采用高斯混合模型进行编码，降低码流，减少图像所占用的存储空间。实验结果表明：在相同码率下SMIC模型的高分六号多光谱图像压缩效果明显优于传统图像压缩算法JPEG，重建图像的质量明显提高，图像的峰值信噪比较JPEG高约2 dB，且SMIC重建图像的误差值主要集中在[-100, 100]范围内，区间占比达到80%以上；SMIC模型重建图像的NDVI与原始图像NDVI的决定系数$ {R}^{2} $为0.93；SMIC模型的冬小麦提取准确率为87.16%，误检率为4.47%，冬小麦提取结果验证了SMIC模型能够满足部分定量遥感的应用需求。

关键词: 图像压缩, 卷积神经网络, 高分六号, 多光谱图像, 自编码器

Abstract:

The spatial and inter-spectral redundancy of the multispectral images of Gaofen No.6 (GF-6) are high.To reduce the storage space occupied by these images and improve their compression efficiency, an end-to-end multispectral image compression model SMIC based on a convolutional neural network is proposed.The SMIC model consists of three parts: self encoder, quantization structure, and entropy coding.The self encoder extracts image features through convolutional down sampling to reduce the spatial dimension of the data.The quantization structure uses the multi-band quantization structure to discretize the feature matrix, reducing information loss in the image compression process.The entropy coding adopts the Gaussian mixture model to reduce the code stream and storage space occupied by the image.Experimental results at the same bit rate show that the compression effect of the SMIC model's GF-6 multispectral images is significantly better than that of the traditional JPEG image compression algorithm, and the quality of the reconstructed image is significantly improved.The Peak Signal-to-Noise Ratio (PSNR) of the image is approximately 2 dB higher than that of JPEG, and the error value of the reconstructed image is mostly within [-100, 100], accounting for more than 80%.The determination coefficient R² between the Normalized Difference Vegetation Index (NDVI) extracted from the reconstructed image and NDVI of the original image is 0.93.The quantitative remote sensing results of a winter wheat extraction experiment show an SMIC model accuracy of 87.16%, and the false detection rate is 4.47%.The results of winter wheat extraction show that the SMIC model can meet the application requirements of partial quantitative remote sensing.

Key words: image compression, convolutional neural network, Gaofen No.6(GF-6), multispectral image, autoencoder

朱孟栩, 张文豪, 李国洪, 顾行发, 余涛, 郑逢杰, 张丽丽, 吴俣, 邴芳飞, 唐健雄. 基于卷积神经网络的高分六号卫星多光谱图像压缩[J]. 计算机工程, 2023, 49(9): 287-294.

Mengxu ZHU, Wenhao ZHANG, Guohong LI, Xingfa GU, Tao YU, Fengjie ZHENG, Lili ZHANG, Yu WU, Fangfei BING, Jianxiong TANG. GF-6 Multispectral Image Compression Based on Convolutional Neural Network[J]. Computer Engineering, 2023, 49(9): 287-294.

http://www.ecice06.com/CN/Y2023/V49/I9/287

图/表 10

图1 数据预处理流程

Fig.1 Data preprocessing procedure

图2 SMIC模型整体技术路线

Fig.2 Overall technique flowchart of SMIC model

图3 SMIC模型结构

Fig.3 Structure of SMIC model

图4 模型训练信息损失曲线

Fig.4 Loss curve of model training information

图5 不同码率下的测试数据PSNR值

Fig.5 PSNR values of test data under different code rates

图6 重建图像比较

Fig.6 Reconstruction images comparison

图7 不同压缩算法重建图像的DN值对比

Fig.7 Comparison of DN values of reconstructed images using different compression algorithms

图8 不同压缩算法重建图像的误差统计直方图

Fig.8 Error statistical histograms of reconstructed images using different compression algorithms

图9 归一化植被指数和冬小麦反演结果

Fig.9 Normalized vegetation indexes and inversion results of winter wheat

图10 NDVI散点图对比

Fig.10 NDVI scatter diagram comparison

参考文献 25

1	刘宝康, 王仁军, 尤晓妮, 等. 基于高分六号WFV数据的冬小麦种植面积提取. 测绘与空间地理信息, 2021, 44(1): 1- 4.
	LIU B K, WANG R J, YOU X N, et al. Extraction of winter wheat area based on GF6-WFV remote sensing image. Geomatics & Spatial Information Technology, 2021, 44(1): 1- 4.
2	MIELIKAINEN J, TOIVANEN P. Lossless compression of hyperspectral images using a quantized index to lookup tables. IEEE Geoscience and Remote Sensing Letters, 2008, 5(3): 474- 478. doi: 10.1109/LGRS.2008.917598
3	RYAN M J, ARNOLD J F. The lossless compression of AVIRIS images by vector quantization. IEEE Transactions on Geoscience and Remote Sensing, 1997, 35(3): 546- 550. doi: 10.1109/36.581964
4	LIU X M, WU X L, ZHOU J T, et al. Data-driven soft decoding of compressed images in dual transform-pixel domain. IEEE Transactions on Image Processing, 2016, 25(4): 1649- 1659. doi: 10.1109/TIP.2016.2526910
5	于恒, 梅红岩, 许晓明, 等. 基于深度学习的图像压缩算法研究综述. 计算机工程与应用, 2020, 56(15): 15- 23. URL
	YU H, MEI H Y, XU X M, et al. Survey of image compression algorithm based on deep learning. Computer Engineering and Applications, 2020, 56(15): 15- 23. URL
6	BALLÉ J, LAPARRA V, SIMONCELLI E P. End-to-end optimized image compression[EB/OL]. [2022-04-05]. https://arxiv.org/abs/1611.01704.
7	BALLÉ J, LAPARRA V, SIMONCELLI E P. End-to-end optimization of nonlinear transform codes for perceptual quality[C]//Proceedings of Picture Coding Symposium. Washington D. C., USA: IEEE Press, 2017: 1-5.
8	BALLÉ J, MINNEN D, SINGH S, et al. Variational image compression with a scale hyperprior[EB/OL]. [2022-04-05]. https://arxiv.org/abs/1802.01436.
9	LIU J, LU G, HU Z, et al. A unified end-to-end framework for efficient deep image compression[EB/OL]. [2022-04-05]. https://arxiv.org/abs/2002.03370.
10	ONG F Q, HU K D, LI Y S, et al. Spectral-spatial feature partitioned extraction based on CNN for multispectral image compression. Remote Sensing, 2020, 13(1): 9. doi: 10.3390/rs13010009
11	刘东, 王叶斐, 林建平, 等. 端到端优化的图像压缩技术进展. 计算机科学, 2021, 48(3): 1- 8. URL
	LIU D, WANG Y F, LIN J P, et al. Advances in end-to-end optimized image compression technologies. Computer Science, 2021, 48(3): 1- 8. URL
12	KONG F, ZHAO S, LI Y, et al. A residual network framework based on weighted feature channels for multispectral image compression. Ad Hoc Networks, 2020, 107, 102272. doi: 10.1016/j.adhoc.2020.102272
13	LI M, ZUO W M, GU S H, et al. Learning convolutional networks for content-weighted image compression[C]//Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington D. C., USA: IEEE Press, 2018: 3214-3223.
14	MENTZER F, AGUSTSSON E, TSCHANNEN M, et al. Conditional probability models for deep image compression[C]//Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington D. C., USA: IEEE Press, 2018: 4394-4402.
15	BALLÉ J, LAPARRA V, SIMONCELLI E P. Density modeling of images using a generalized normalization transformation[EB/OL]. [2022-04-05]. https://arxiv.org/abs/1511.06281.
16	THEIS L, SHI W, CUNNINGHAM A, et al. Lossy image compression with compressive autoencoders[EB/OL]. [2022-04-05]. https://arxiv.org/abs/1703.00395.
17	王铭, 姜淑华, 吴杰, 等. 基于改进生成式对抗网络的图像去雾算法研究. 长春理工大学学报(自然科学版), 2021, 44(2): 93- 99. URL
	WANG M, JIANG S H, WU J, et al. Research on image defogging algorithm based on improved generative antagonistic network. Journal of Changchun University of Science and Technology(Netural Science Edition), 2021, 44(2): 93- 99. URL
18	MINNEN D, BALLÉ J, TODERICI G. Joint autoregressive and hierarchical priors for learned image compression[C]//Proceedings of the 32nd International Conference on Neural Information Processing Systems. New York, USA: ACM Press, 2018: 10794-10803.
19	LEE J, CHO S, KIM M. An end-to-end joint learning scheme of image compression and quality enhancement with improved entropy minimization[EB/OL]. [2022-04-05]. https://arxiv.org/abs/1912.12817.
20	YANG Y H. Elements of information theory. Journal of the American Statistical Association, 2008, 103(481): 429.
21	方知. 高光谱图像降维压缩比自动检测数学模型仿真. 计算机仿真, 2021, 38(4): 487- 491. URL
	FANG Z. Mathematical model simulation of dimension reduction compression ratio automatic detection for hyperspectral images. Computer Simulation, 2021, 38(4): 487- 491. URL
22	YD A, RSS A, KP B, et al. Convolution neural network based lossy compression of hyperspectral images. Signal Processing: Image Communication, 2021, 95, 116255.
23	WALLACE G K. The JPEG still picture compression standard. IEEE Transactions on Consumer Electronics, 1992, 38(1): 1- 17.
24	梁继, 郑镇炜, 夏诗婷, 等. 高分六号红边特征的农作物识别与评估. 遥感学报, 2020, 24(10): 1168- 1179.
	LIANG J, ZHENG Z W, XIA S T, et al. Crop recognition and evaluationusing red edge features of GF-6 satellite. Journal of Remote Sensing, 2020, 24(10): 1168- 1179.
25	赵兵杰. 基于GF-1与Landsat-8的种植结构提取与产量估测[D]. 邯郸: 河北工程大学, 2019.
	ZHAO B J. Extraction of planting structure and yield estimation based on GF-1 and Landsat-8[D]. Handan: Hebei University of Engineering, 2019. (in Chinese)

[1]	李现国, 李滨. 基于Transformer和多尺度CNN的图像去模糊[J]. 计算机工程, 2023, 49(9): 226-233, 245.
[2]	杜逸潇, 王红军, 李修和. 基于频谱地图的辐射源指纹定位方法研究[J]. 计算机工程, 2023, 49(9): 183-190, 198.
[3]	胡水. 基于深度强化学习的智能兵棋推演决策方法[J]. 计算机工程, 2023, 49(9): 303-312.
[4]	韩璐, 霍纬纲, 张永会, 刘涛. 基于多尺度特征融合与双注意力机制的多元时间序列预测[J]. 计算机工程, 2023, 49(9): 99-108.
[5]	李哲铭, 王晋东, 侯建中, 李伟, 张世华, 张恒巍. 基于显著区域优化的对抗样本攻击方法[J]. 计算机工程, 2023, 49(9): 246-255, 264.
[6]	曹坪, 杨怀志, 薄一军, 尤嘉, 张淳杰, 李丹勇. 面向低质量裂缝图像的多知识蒸馏分类[J]. 计算机工程, 2023, 49(7): 204-213.
[7]	白明昌. 基于折叠路径聚合的属性网络节点嵌入方法[J]. 计算机工程, 2023, 49(7): 76-84.
[8]	代祖华, 刘园园, 狄世龙. 语义增强的图神经网络方面级文本情感分析[J]. 计算机工程, 2023, 49(6): 71-80.
[9]	沈学利, 田桂源, 姜彦吉, 马琳琳. 基于双阶段Conv-Transformer的时频域语音增强算法[J]. 计算机工程, 2023, 49(6): 123-130.
[10]	丁子轩, 俞雷, 张娟, 李想, 王新宇. 基于深度残差自适应注意力网络的图像超分辨率重建[J]. 计算机工程, 2023, 49(5): 231-238.
[11]	张子宣, 宗学军, 何戡, 连莲. 基于CVAE-CatBoost的工业控制网络异常流量检测研究[J]. 计算机工程, 2023, 49(5): 173-180.
[12]	陈治旭, 靳雁霞, 芦烨, 杨晶, 刘亚变, 史志儒. 基于子图卷积神经网络的多精度服装建模方法[J]. 计算机工程, 2023, 49(4): 174-181.
[13]	徐康, 李霏, 姬东鸿. 结合依存图卷积与文本片段搜索的方面情感三元组抽取[J]. 计算机工程, 2023, 49(4): 61-67.
[14]	衡红军, 苗菁. 语义与句法信息加强的二元标记实体关系联合抽取[J]. 计算机工程, 2023, 49(4): 77-84.
[15]	钟宝荣, 吴夏灵. 基于高分辨率网络的轻量型人体姿态估计研究[J]. 计算机工程, 2023, 49(4): 226-232,239.

选择文件类型/文献管理软件名称

选择包含的内容

基于卷积神经网络的高分六号卫星多光谱图像压缩

GF-6 Multispectral Image Compression Based on Convolutional Neural Network

RichHTML

PDF

可视化

被引次数

摘要/Abstract

引用本文

使用本文

图/表 10

参考文献 25

相关文章 15

编辑推荐

Metrics

本文评价

模态框（Modal）标题

选择文件类型/文献管理软件名称

选择包含的内容

基于卷积神经网络的高分六号卫星多光谱图像压缩

GF-6 Multispectral Image Compression Based on Convolutional Neural Network

RichHTML

PDF

可视化

被引次数

摘要/Abstract

引用本文

使用本文

图/表 10

参考文献 25

相关文章 15

编辑推荐

Metrics

本文评价