-
摘要: 为了实现高像质相机低成本、小型化的需求，本文提出了一种大视场简单光学系统的光学-算法协同设计方法，并通过图像复原算法校正简单光学系统的残余像差。首先，针对大视场光学系统，对空间变化的交叉通道去卷积算法进行改进，加入倍率色差校正，使图像复原算法可显著去除色差的影响。然后，在光学设计过程中，放开色差的约束，并专注优化绿色通道的像质，使其成像锐利，在后期交叉通道去卷积算法中有助于红、蓝两通道图像复原。利用该方法设计了一个由两片同种材料的镜片构成的大视场简单光学系统。系统焦距为50 mm，全视场为46°，F数为5.6，探测器分辨率为1 000万像素。实验结果表明：本文设计的两片镜、大视场简单光学系统的成像质量可媲美三片式库克镜头，明显优于纯图像复原的结果。本文方法实现了大视场简单光学系统的设计，并能够通过系统最终获得高分辨率、高像质图像。Abstract: In order to meet requirements of the low cost and miniaturization of high-quality cameras, an optical/algorithmic co-design method for a large-field simple optical system is proposed. The residual aberrations of this simple optical system are corrected using an image restoration algorithm. Firstly, spatially varying cross-channel deconvolution is improved for large-field optical systems, wherein correction of lateral chromatic aberrations is introduced so that they can be significantly removed through the image restoration algorithm. Then, when designing the optical system, the constraints of chromatic aberrations are removed and we concentrate on optimizing the quality of the green channel to make the image sharp, which will help the restoration of images in the red and the blue channels in later cross-channel deconvolution. Using this method, a large-field simple optical system with two lenses of the same material is designed. The focal length of the system is 50 mm, the full field is 46 degrees, the F number is 5.6 and the resolution of the sensor is 10 mega pixels. Experimental results indicate that the image quality of the proposed simple large-field optical system is comparable to that when using a Cooke triplet lens and is better than that when using pure image restoration algorithms. Our proposed method succeeds as a simple large-field optical system and can obtain high-resolution and high-quality images.
-
Key words:
- computational imaging /
- optical system design /
- image restoration /
- deconvolution
-
表 1 简单光学系统的镜头参数
Table 1. Lens data of the simple optical system
Surface Radius Thickness Glass Semi-Aperture Object Infinity Infinity 1 7.872 8 1.71 HK9L_CDGM 6.30 2 6.582 4 5.57 5.38 stop Infinity 4.28 4.28 4 -33.333 4 10.00 HK9L_CDGM 6.28 5 -13.943 6 56.45 9.10 Image Infinity 0 20.21 
下载: 导出CSV
表 2 BRISQUE评价结果对比
Table 2. Comparison of BRISQUE scores
Flowers Campus Ground truth 17.079 6 21.843 0 Krishnan et al. 51.551 5 49.697 0 Heide et al. 44.991 0 44.785 0 Ours 30.548 2 39.275 0 Cooke lens 29.490 0 36.888 0
下载: 导出CSV
表 3 NIQE评价结果对比
Table 3. Comparison of NIQE scores
Flowers Campus Ground truth 3.238 0 3.333 8 Krishnan et al. 6.612 2 6.527 4 Heide et al. 6.059 0 5.825 9 Ours 4.838 0 4.577 5 Cooke lens 4.200 0 4.413 1
下载: 导出CSV
-
[1] 杨利红, 赵变红, 张星祥, 等.点扩散函数高斯拟合估计与遥感图像恢复[J].中国光学, 2012, 5(2):181-188. http://www.chineseoptics.net.cn/CN/abstract/abstract8812.shtmlYANG L H, ZHAO B H, ZHANG X X, et al.. Gaussian fitted estimation of point spread function and remote sensing image restoration[J]. Chinese Optics, 2012, 5(2):181-188.(in Chinese) http://www.chineseoptics.net.cn/CN/abstract/abstract8812.shtml [2] 闫敬文, 彭鸿, 刘蕾, 等.基于L0正则化模糊核估计的遥感图像复原[J].光学 精密工程, 2014, 22(9):2572-2579. http://d.old.wanfangdata.com.cn/Periodical/gxjmgc201409037YAN J W, PENG H, LIU L, et al.. Remote sensing image restoration based on zero-norm regularized kernel estimation[J]. Opt. Precision Eng., 2014, 22(9):2572-2579.(in Chinese) http://d.old.wanfangdata.com.cn/Periodical/gxjmgc201409037 [3] 曹海霞, 巴音贺希格, 崔继承, 等.基于图像复原的高光谱图像前向像移补偿[J].中国光学, 2013, 6(6):856-862. http://www.chineseoptics.net.cn/CN/abstract/abstract8994.shtmlCAO H X, BAYAIHESHIG, CUI J CH, et al.. Forward image motion compensation of hyper spectral image based on image restoration[J]. Chinese Optics, 2013, 6(6):856-862.(in Chinese) http://www.chineseoptics.net.cn/CN/abstract/abstract8994.shtml [4] 杨航, 吴笑天, 王宇庆.基于结构字典学习的图像复原方法[J].中国光学, 2017, 10(2):207-218. http://d.old.wanfangdata.com.cn/Periodical/zggxyyygxwz201702005YANG H, WU X T, WANG Y Q. Image restoration approach based on structure dictionary learning[J]. Chinese Optics, 2017, 10(2):207-218.(in Chinese) http://d.old.wanfangdata.com.cn/Periodical/zggxyyygxwz201702005 [5] WIENER N. Extrapolation, Interpolation, and Smoothing of Stationary Time Series[M]. Cambridge, Massachusetts, USA:MIT Press, 1964. [6] RICHARDSON W H. Bayesian-based iterative method of image restoration[J]. Journal of the Optical Society of America, 1972, 62(1):55-59. doi: 10.1364/JOSA.62.000055 [7] LUCY L B. An iterative technique for the rectification of observed distributions[J]. The Astronomical Journal, 1974, 79(6):745-754. doi: 10.1086/111605 [8] LEVIN A, FERGUS R, DURAND F, et al.. Image and depth from a conventional camera with a coded aperture[J]. ACM Transactions on Graphics, 2007, 26(3):70. doi: 10.1145/1276377.1276464 [9] KRISHNAN D, FERGUS R. Fast image deconvolution using hyper-Laplacian priors[C]. Advances in Neural Information Processing Systems 22: 23rd Annual Conference on Neural Information Processing Systems 2009, NIPS, 2009, 22: 1-9. [10] 郭从洲, 秦志远.非凸高阶全变差正则化自然光学图像盲复原[J].光学 精密工程, 2015, 23(12):3490-3499. http://d.old.wanfangdata.com.cn/Periodical/gxjmgc201512027GUO C ZH, QIN ZH Y. Blind restoration of nature optical images based on non-convex high order total variation regularization[J]. Opt. Precision Eng., 2015, 23(12):3490-3499.(in Chinese) http://d.old.wanfangdata.com.cn/Periodical/gxjmgc201512027 [11] YUE T, SUO J L, WANG J, et al.. Blind optical aberration correction by exploring geometric and visual priors[C]. 2015 IEEE Conference on Computer Vision and Pattern Recognition, IEEE, 2015: 1684-1692. [12] 张金刚, 相里斌, 汶德胜, 等.采用色差先验约束的像差校正技术[J].中国光学, 2018, 11(4):560-567. http://www.chineseoptics.net.cn/CN/abstract/abstract9579.shtmlZHANG J G, XIANG L B, WEN D SH, et al.. Aberration correction technology based on chromatic aberration prior constraints[J]. Chinese Optics, 2018, 11(4):560-567.(in Chinese) http://www.chineseoptics.net.cn/CN/abstract/abstract9579.shtml [13] SCHULER C J, HIRSCH M, HARMELING S, et al.. Non-stationary correction of optical aberrations[C]. 2011 International Conference on Computer Vision, IEEE, 2011: 659-666. [14] HEIDE F, ROUF M, HULLIN M B, et al.. High-quality computational imaging through simple lenses[J]. ACM Transactions on Graphics, 2013, 32(5):149. http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=fbb2ee3c18e52a4fff4d6d4f6cea3b40 [15] LI W L, LIU Y, YIN X Q, et al.. A computational photography algorithm for quality enhancement of single lens imaging deblurring[J]. Optik, 2015, 126(21):2788-2792. doi: 10.1016/j.ijleo.2015.07.030 [16] WANG J Y, WANG L, YANG Y, et al.. An integral design strategy combining optical system and image processing to obtain high resolution images[J]. Proceedings of SPIE, 2016, 9874:98740F. doi: 10.1117/12.2224394 [17] 谭政, 相里斌, 吕群波, 等.基于像差选择性校正的光学-数字联合设计[J].光子学报, 2018, 47(5):0511001. http://d.old.wanfangdata.com.cn/Periodical/gzxb201805022TAN ZH, XIANG L B, LV Q B, et al.. Optics/digital processing co-design method based on aberration optional-correcting[J]. Acta Photonica Sinica, 2018, 47(5):0511001. in Chinese http://d.old.wanfangdata.com.cn/Periodical/gzxb201805022 [18] 崔金林.基于计算光学的非完善光学系统图像质量提高及其应用研究[D].长春: 中国科学院长春光学精密机械与物理研究所, 2018.CUI J L. Research on enhancing imaging quality of imperfect optical systems and its application via computational optics[D]. Changchun: Changchun Institute of Optics, Fine Mechanics and Physics, Chinese Academy of Sciences, 2018.(in Chinese) [19] CHAMBOLLE A, POCK T. A first-order primal-dual algorithm for convex problems with applications to imaging[J]. Journal of Mathematical Imaging and Vision, 2011, 40(1):120-145. doi: 10.1007/s10851-010-0251-1 [20] 张以谟.应用光学[M].北京:电子工业出版社, 2012.ZHANG Y M. Applied Optics[M]. Beijing:Publishing House of Electronics Industry, 2012.(in Chinese) [21] MITTAL A, MOORTHY A K, BOVIK A C. No-reference image quality assessment in the spatial domain[J]. IEEE Transactions on Image Processing, 2012, 21(12):4695-4708. doi: 10.1109/TIP.2012.2214050 [22] MITTAL A, SOUNDARARAJAN R, BOVIK A C. Making a "completely blind" image quality analyzer[J]. IEEE Signal Processing Letters, 2013, 20(3):209-212. doi: 10.1109/LSP.2012.2227726 -
下载:
点击查看大图
图(11) / 表(3)
计量
- 文章访问数: 1456
- HTML全文浏览量: 466
- PDF下载量: 60
- 被引次数: 0
