基于Res-CAE深度学习实现高光谱图谱融合的国画重建

doi:10.3964/j.issn.1000-0593(2025)09-2428-09

摘要
参考文献
相关文章 (15)

全文: PDF (27454 KB)
输出: BibTeX | EndNote (RIS)

摘要：针对传统颜色再现方法存在实验数据预处理复杂且依赖人为主观光谱特征选择、光谱反射率数据空间信息利用缺失的问题，本研究提出基于深度学习的高光谱数据立方体协同分析方法。通过构建残差-卷积自编码器(Res-CAE)深度网络模型，实现对高光谱数据空间维度与光谱维度的多维信息协同学习与联合重建。选择CAVE高光谱数据集作为Res-CAE的训练集数据，而后根据实验需要选择爱色丽24色标准色卡、自制中国画色卡、训练集中随机场景、训练集外随机场景和某国画在国际照明委员会标准观察者环境下的高光谱数据这五个场景作为测试集数据来对Res-CAE的重建效果进行测试。对于测试结果在光谱方面使用色差、均方根差进行评价，在图像质量方面使用峰值信噪比、结构相似性指数等评价标准进行评价，同时将Res-CAE的实验结果与双线性插值法、主成分分析法和Res-CAE的消融实验进行实验结果对比。从实验分析可得出Res-CAE在色差、重建图象质量、重建时间等方面要好于其他方法。以爱色丽24色标准色卡为例，Res-CAE的重建平均色差为0.694 5，平均均方根差为0.009 2，平均峰值信噪比为35.923 0，结构相似性指数为0.995 6。实验证明，Res-CAE可以很好的应用在颜色重建方面。

关键词：高光谱数据；颜色重建；光谱反射率重建；深度学习

Abstract：Traditional color reproduction methods often suffer from complex preprocessing steps and reliance on subjective selection of spectral features. Moreover, the exclusive use of spectral reflectance data neglects spatial information, limiting reconstruction to isolated color points rather than full scenes. To overcome these limitations, this study proposes a deep learningbased method using a Residual-Convolutional Autoencoder (Res-CAE) to jointly extract and reconstruct spatial and spectral features from hyperspectral data cubes. The Res-CAE model was trained on the CAVE hyperspectral dataset and evaluated across five testing scenarios: a standard 24-color chart (X-Rite), a custom Chinese painting color chart, in-training and out-of-training random scenes, and a real Chinese painting scene captured under CIE standard observer conditions. Evaluation metrics included color difference (ΔE₀₀), root mean square error (RMSE), peak signal-to-noise ratio (PSNR), and structural similarity index (SSIM). Experimental results demonstrate that Res-CAE outperforms traditional methods, such as bilinear interpolation and principal component analysis (PCA), in both color fidelity and image quality. On the 24-color chart, the model achieved an average ΔE₀₀ of 0.694 5, RMSE of 0.009 2, PSNR of 35.92, and SSIM of 0.995 6. These results validate the effectiveness of Res-CAE in high-fidelity color reconstruction from hyperspectral data, offering practical value for digital preservation of traditional Chinese paintings.

Key words：Hyperspectral data; Colorrecon struction; Spectral reflectance reconstruction; Deep learning

收稿日期: 2024-11-25 修订日期: 2025-05-29

中图分类号:

O433

基金资助: 云南省廖桂生专家工作站（202305AF150012)，国家自然科学基金项目（62065019），云南师范大学研究生科研创新基金项目（YJSJJ22-A18），云南师范大学2024年度研究生科研创新基金项目（YISII23-B70）资助

通讯作者: 冯洁 E-mail: fengjie_ynnu@126.com

作者简介: 朱世豪，1997年生，云南师范大学物理与电子信息学院硕士研究生 e-mail: 377938599@qq.com

引用本文:

朱世豪，冯洁，李欣庭，孙丽存，刘洁，袁娉，杨仁翔，邓红杨. 基于Res-CAE深度学习实现高光谱图谱融合的国画重建[J]. 光谱学与光谱分析, 2025, 45(09): 2428-2436.
ZHU Shi-hao, FENG Jie, LI Xin-ting, SUN Li-cun, LIU Jie, YUAN Ping, YANG Ren-xiang, DENG Hong-yang. Reconstruction of Chinese Paintings Based on Hyperspectral Image Fusion Using Res-CAE Deep Learning. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2025, 45(09): 2428-2436.

链接本文:

https://www.gpxygpfx.com/CN/10.3964/j.issn.1000-0593(2025)09-2428-09 或 https://www.gpxygpfx.com/CN/Y2025/V45/I09/2428

[1] ZHOU Jie, DUAN Lian-ru(周杰，段炼孺). Identification and Appreciation to Cultural Relics(文物鉴定与鉴赏), 2021, 198: 78.
[2] WANG Hong-mei(王红梅). Oriental Collection(东方收藏), 2021, (13): 76.
[3] XU Guo-bin, YU Yi-ming, LI Jie, et al(徐国彬，于艺铭，李洁, 等). Laser & Optoelectronics Progress, 2022, 59(10): 81.
[4] CHENG Qing-biao, CHEN Guang-yun, WANG Da-wen, et al(程青彪, 陈广云, 王大文, 等). Optical Instruments（光学仪器）, 2022, 44(3): 68.
[5] Zhang L H, Liang D, Zhang D W, et al. Journal of the Optical Society of Korea, 2016, 20(4): 515.
[6] Wang K, Wang H, Wang Z, et al. Cluster Computing, 2019, 22(S1): 493.
[7] YAN Li-xia, WU Fan(闫丽霞, 吴凡). Computer Applications and Software(计算机应用与软件), 2014, 31(3): 207.
[8] ZHAO Hai, LI Hong-ning, CHEN Hao(赵海, 李宏宁, 陈豪). Acta Optica Sinica(光学学报), 2023, 43(9): 269.
[9] WANG Wen-ju, WANG Jiang-wei(王文举, 王江薇). Packaging Engineering(包装工程), 2020, 41(11): 254.
[10] Alvarez Gila A, Van de Weijer J, Garrote E. Proceedings of the 2017 IEEE International Conference on Computer Vision Workshops, 2017, 480.
[11] Arad B, Ben-Shahar O. Computer Vision—14th European Conference, ECCV 2016 Proceedings, PartVII14, Springer: 2016, 19.
[12] Cai Y, Lin J, Hu X, et al. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022, 17502.
[13] Cai Y, Lin J, Hu X, et al. Proceedings European Conference on Computer Vision, Springer, 2022, 686.
[14] Cai Y, Lin J, Lin Z, et al. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022, 745.
[15] Urban P, Rosen M R, Berns R S. Journal of the Optical Society of America A, 2009, 26(8): 1865.
[16] LIU Peng-fei, ZHAO Huai-ci, LI Pei-xuan(刘鹏飞, 赵怀慈, 李培玄). Infrared and Laser Engineering(红外与激光工程), 2020, 49(S1): 143.
[17] LI Xin-ting, ZHANG Feng, FENG Jie(李欣庭, 张峰, 冯洁). Spectroscopy and Spectral Analysis(光谱学与光谱分析), 2024, 44(1): 215.
[18] Li Y, Hu J, Zhao X, et al. Neurocomputing, 2017, 266: 29.
[19] Shi Z, Chen C, Xiong Z, et al. Proceedings of the IEEE Conference on Computer Visionand Pattern Recognition Workshops, 2018, 939.
[20] Hu X, Cai Y, Lin J, et al. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022, 17542.
[21] Huang T, Dong W, Yuan X, et al. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2021, 16216.