窄带快照式多光谱成像色彩还原方法

doi:10.3964/j.issn.1000-0593(2021)01-0183-05

摘要
参考文献
相关文章 (1)

全文: PDF (2195 KB)
输出: BibTeX | EndNote (RIS)

摘要：由于兼具光谱分辨和空间分辨能力, 快照式窄带多光谱成像在资源遥感、精准农业、医疗检测等领域将有广泛应用。由于该方法使用窄带成像来提高光谱分辨率及图像对比度，所获得的图像是灰度信息，失去了场景的色彩信息，不便专家对图像鉴别、评价与赏析。已有的色彩还原算法主要针对光谱波段带宽较宽或者多个波段叠加基本覆盖整个可见光谱范围等两种光谱成像仪，不适合窄带多光谱成像方法的色彩还原。该研究适合于快照式窄带多光谱成像的色彩还原方法，提出建立窄带多光谱彩色相机的概念。首先，提出两种窄带多光谱色彩还原方法：(1)基于CIE色度系统三刺激值色的，(2)基于贝尔阵列插值算法的；其次，分别应用两种算法还原快照式窄带多光谱相机所获得的植物、手臂及宫颈组织等三种代表性场景窄带多光谱灰度图像；之后，计算并比较表征两种算法所得的代表性场景彩色图像的均值、方差、熵及梯度等表征图像质量的参数数值，确定出适合快照式窄带多光谱成像的色彩还原方法；最后，对所确定的色彩还原算法进行色偏校正。实验结果表明，基于CIE三刺激值色彩还原方法比贝尔阵列插值法更适用于窄带多光谱成像颜色复原。配合使用CIE三刺激值色彩还原方法及灰度图象校正算法，从窄带光谱成像所获得的植物、手臂皮肤及宫颈组织的灰度图像所还原出的近彩色图像逼近物体真实色彩，满足人眼观察习惯。介绍了仅覆盖可见光光谱范围30%的窄带多光谱图像进行色彩还原的方法，该方法证明快照式窄带多光谱成像可以兼具光谱分辨能力，同时保持可供人主观辨识的色彩信息。所提出实现快照式窄带多光谱彩色成像的方法，有望设计不同于传统RGB相机的彩色相机实施方案。

关键词：快照式窄带多光谱成像；色彩还原方法；CIE三刺激值积分；贝尔阵列插值；窄带多光谱彩色相机

Abstract：Due to its simultaneous spectral and spatial resolutions, snapshot narrow-band spectral imaging (SNBSI) will have wide applications in remote sensing of natural resources, accurate agriculture, and medical diagnosis. However, due to its usage of narrow bands to improve spectral resolution and to enhance contrast, gray images instead of colorful ones produced by SNBSI are not idea for experts to identify, evaluate and appreciate. Existing color recovery methods primarily applicable to either wide-band spectral imaging or continuous spectral imaging wherein multiple bands together cover the full visible spectral region; however, they are not directly applicable to SNBSI. In this paper, we study color recovery method for SNBSI. Further we propose the concept of color camera via narrow-band multi-spectral imaging method. First, two color recovery methods were proposed, one was based on CIE three primary coordinate system; the other was modified Bayer-filter-array based color interpolation method. Then applied the two proposed methods to resurrect color informationfor three representative targets including a plant, a forearm, and a piece of cervical tissues captured by an SNBSI camera, then selected one of the two methods which resulted in higher values of mean, standard deviation, entropy, and gradient of the recovered images. Finally, the modified gray world chromatic aberration correction algorithm was applied to make the recovered images closer to the true color. Experimental results suggested that CIE-three primary based method was more suitable than the Bayer-filter-array based interpolation method for color recovery from SNBI spectral images. The recovered color images of a plant, a human forearm, and tissues by the CIE three primary based recovery method followed by the modified gray world chromatic correction algorithm were close to the true color and were satisfactory for human observation. This study introduces acolor recovery method which exploits few narrow-band spectral images that jointly cover about 30% of the visible spectral range. The existence of such a color recovery method proves that SNBSI is capable of having both spectral resolution and retain color information. Therefore, this paper proposes a method to realize snapshot narrow-band multi-spectral color imaging, which opens up a color restoration method different from the conventional RGB color camera.

Key words：Snapshot narrow-band spectral imaging；Color recovery method；CIE three primary color；Bayer filter array interpolation; Narrow band spectral imaging color camera

收稿日期: 2019-03-26 修订日期: 2020-02-01

中图分类号:

TP73

基金资助: 国家自然科学基金项目（51775200），数字福建工业制造物联网实验室项目(KY050004)资助

通讯作者: 孔令华 E-mail: klh@fjut.edu.cn

作者简介: 易定容，女，1969年生，华侨大学机电及自动化学院教授 e-mail: yidr@hqu.edu.cn

引用本文:

易定容，孔令华，赵艳丽，杨子涵. 窄带快照式多光谱成像色彩还原方法[J]. 光谱学与光谱分析, 2021, 41(01): 183-187.
YI Ding-rong, KONG Ling-hua, ZHAO Yan-li, YANG Zi-han. Color Recovery Method for Snapshot Narrow Band Spectral Imaging Technology. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2021, 41(01): 183-187.

链接本文:

https://www.gpxygpfx.com/CN/10.3964/j.issn.1000-0593(2021)01-0183-05 或 https://www.gpxygpfx.com/CN/Y2021/V41/I01/183

[1] Shen Y, Lei Y U, Jiexiang C. Acta Photonica Sinica，2018，47(8)：822003.
[2] Ren D D W, Tripathi S. Journal of Applied Remote Sensing，2017，11：016006.
[3] Zapata F, Lópezlópez M, Amigo J M. Multi-Spectral Imaging for the Estimation of Shooting Distances，2018，282：80.
[4] Estelleslopez L，Ropodi A, Pavlidis, et al. Food Research International，2017，99(1)：206.
[5] Valm A M，Cohen S, Legant W R, et al. Nature，2017，546：162.
[6] Yi Dingrong，Wang Chao，Qi Hairong. IEEE Transactions on Biomedical Engineering, 2011，58(3): 736.
[7] YI Ding-rong, KONG Ling-hua, LIU Ting(易定容，孔令华，刘婷). Infrared and Laser Engineering(红外与激光工程)，2017，46(7): 256.
[8] Zhang C M, Wu H, Li J. Optical Engineering，2011，50(6): 0600201.
[9] WANG Xin-quan, XIANGLI Bin, HUANG Min(王新全, 相里斌, 黄旻). Journal of Optoelectronics·Laser(光电子·激光)，2011，5(3)：689.
[10] ZHANG Yu, ZHU Shuai-shuai, ZHAO Liang-yu(张宇，朱帅帅，赵梁玉). Journal of Harbin Institute of Technology(哈尔滨工业大学学报)，2017，49(3): 29.
[11] Kong L，Yi D, Sprigle S, et al. Journal of Biomedical Optics，2010，15(1): 502.
[12] HUANG Hong-lian, YI Wei-ning, DU Li-li(黄红莲，易维宁，杜丽丽). Infrared and Laser Engineering(红外与激光工程)，2016，45(11): 354.
[13] LÜ Xiao-kai, LI Chuan-rong, MA Ling-ling(吕晓凯，李传荣，马灵玲). Remote Sensing Technology and Application(遥感技术与应用)，2013，28(3): 467.
[14] HE Song-hua, LIU Zhen, CHEN Qiao(何颂华, 刘真, 陈桥). Acta Optical Sinica(光学学报)，2014，34：325.
[15] WANG Jia-jia，LIAO Ning-fang, WU Wen-min, et al(王佳佳，廖宁放，吴文敏，等). Spectroscopy and Spectral Analysis(光谱学与光谱分析)，2017，37(3)：704.
[16] XU Hui-xi, CHEN Yun-hao, XUE Wan-rong(许辉熙, 陈云浩, 薛万蓉). Laser and Optoelectronics Progress(激光与光电子学进展)，2015，52(5): 77.