运动模糊图像PSF参数估计方法改进及图像复原

doi:10.19678/j.issn.1000-3428.0062656

摘要/Abstract

摘要： 运动模糊图像复原的目的是改善运动图像质量，从而为图像处理任务提供高质量的清晰图像以保证算法能够准确获取图像信息，其中运动模糊图像的点扩散函数（PSF）求解是影响复原图像质量的关键步骤。针对现有运动模糊图像PSF参数估计方法中存在的估计误差大、有效估计范围有限等问题，在分析频谱图像特征的基础上，提出一种改进的PSF参数估计方法。通过图像增强处理和形态学变换去除频谱图像中的十字亮线和噪点干扰，获取形态合适的条纹图像以完成Radon变换检测。利用二值频谱图像的条纹特征自适应地控制形态学运算精度，从而保证算法的执行效率和鲁棒性。对条纹进行边缘测定，消除由条纹自身宽度导致的角度估计误差，以提高参数估计结果的精度。实验结果表明，该方法能够提高模糊参数估计的准确率和有效估计范围，由此构建的PSF能复原出更加清晰的重建图像，复原图像总体峰值信噪比不低于25 dB。

关键词: 图像处理, 图像复原, 运动模糊图像, 点扩散函数, 形态学运算, Radon变换

Abstract: The purpose of motion-blurred image restoration involves improving the quality of moving images to provide high-quality clear images for image processing tasks.This ensures that the algorithm can accurately obtain image information.The Point Spread Function(PSF) solution of motion-blurred images is a key step that affects the quality of restored images.Given the problems of large estimation error and limited effective estimation range in the existing PSF parameter estimation methods for motion-blurred images, an improved PSF parameter estimation method is proposed based on the analysis of spectral image features.The cross bright lines and noise interference in the spectrum image are removed via image enhancement and morphological transformation, and the fringe image with appropriate morphology is obtained to complete the Radon transform detection.The fringe feature of binary spectrum image is used to adaptively control the accuracy of morphological operation to ensure the efficiency and robustness of the algorithm.The fringe edge is measured to eliminate the angle estimation error due to the fringe width for improving the accuracy of parameter estimation results.Experimental results showed that this method can improve the accuracy and effective estimation range of fuzzy parameter estimation.The PSF constructed from this method can restore a clearer reconstructed image, and the overall Peak Signal-to-Noise Ratio(PSNR) of the restored image is not less than 25 dB.

Key words: image processing, image restoration, motion-blurred image, Point Spread Function(PSF), morphology operations, Radon transform

中图分类号:

TP391.41

高海韬, 李丹宁, 王彬, 唐鑫鑫. 运动模糊图像PSF参数估计方法改进及图像复原[J]. 计算机工程, 2022, 48(9): 197-203,212.

GAO Haitao, LI Danning, WANG Bin, TANG Xinxin. Improvement of PSF Parameter Estimation Method for Motion-Blurred Image and Image Restoration[J]. Computer Engineering, 2022, 48(9): 197-203,212.

http://www.ecice06.com/CN/Y2022/V48/I9/197

图/表 10

20221008183258

20221008183302

20221008183305

20221008183309

20221008183313

20221008183317

20221008183321

20221008183324

20221008183328

20221008183332

参考文献

[1] ZHANG T T, SONG S S, JIA Z H, et al.Object motion deblurring in single image under static background[J].IEEE Access, 2020, 8:218069-218080.
[2] TU Z G, POPPE R, VELTKAMP R.Estimating accurate optical flow in the presence of motion blur[J].Journal of Electronic Imaging, 2015, 25(5):530181-530189.
[3] MA Z Y, LIAO R J, TAO X, et al.Handling motion blur in multi-frame super-resolution[C]//Proceedings of 2015 IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C., USA:IEEE Press, 2015:5224-5232.
[4] 王伟鹏, 项文杰.结合卷积变换与细节优化的图像去雾算法[J].小型微型计算机系统, 2021, 42(4):842-846. WANG W P, XIANG W J.Image dehazing algorithm based on convolution transformation and detail optimization[J].Journal of Chinese Computer Systems, 2021, 42(4):842-846.(in Chinese)
[5] 缪斯, 祝永新.针对图像盲去模糊的可微分神经网络架构搜索方法[J].计算机工程, 2021, 47(9):313-320. MIAO S, ZHU Y X.Differentiable neural architecture search method for blind image deblurring[J].Computer Engineering, 2021, 47(9):313-320.(in Chinese)
[6] TRULL A K, VAN DER HORST J, PALENSTIJN W J, et al.Point spread function based image reconstruction in optical projection tomography[J].Physics in Medicine and Biology, 2017, 62(19):7784-7797.
[7] KHAN A, YIN H J.Arbitrarily shaped Point Spread Function(PSF) estimation for single image blind deblurring[J].The Visual Computer, 2021, 37(7):1661-1671.
[8] SIZIKOV V S.Estimating the point-spread function from the spectrum of a distorted tomographic image[J].Journal of Optical Technology, 2015, 82(10):655.
[9] LEE J M, LEE J H, PARK K T, et al.Image deblurring based on the estimation of PSF parameters and the post-processing[J].Optik, 2013, 124(15):2224-2228.
[10] ZHU X, ROUBEK F, MILANFAR P.Deconvolving PSFs for a better motion deblurring using multiple images[C]//Proceedings of European Conference on Computer Vision.Berlin, Germany:Springer, 2012:636-647.
[11] MOGHADDAM M E, JAMZAD M.Linear motion blur parameter estimation in noisy images using fuzzy sets and power spectrum[J].EURASIP Journal on Advances in Signal Processing, 2007(1):1-8.
[12] LI Z J, ZHAO Q, GONG W L.Distorted point spread function and image reconstruction for ghost imaging[J].Optics and Lasers in Engineering, 2021, 139:106486-106494.
[13] 廖秋香, 卢在盛, 彭金虎.运动模糊图像PSF参数估计与图像复原研究[J].高技术通讯, 2019, 29(4):338-343. LIAO Q X, LU Z S, PENG J H.Research on PSF parameter estimation and image restoration of motion blurred image[J].Chinese High Technology Letters, 2019, 29(4):338-343.(in Chinese)
[14] 胡发焕, 刘国平, 胡瑢华, 等.一种改进的运动模糊图像参数准确估计方法[J].控制与决策, 2018, 33(12):2177-2183. HU F H, LIU G P, HU R H, et al.Improved approach to motion blur parameters identification[J].Control and Decision, 2018, 33(12):2177-2183.(in Chinese)
[15] 乐翔, 程建, 李民.改进的基于Radon变换的运动模糊图像参数估计方法[J].红外与激光工程, 2011, 40(4):679. LE X, CHENG J, LI M.Improved approach to motion blur identification based on Radon transform[J].Infrared and Laser Engineering, 2011, 40(4):679.(in Chinese)
[16] 胡硕, 张旭光, 吴娜.基于Radon变换的运动模糊方向估计的改进方法[J].高技术通讯, 2015, 25(S1):822-828. HU S, ZHANG X G, WU N.Improved identification of the motion blur direction based on Radon transform[J].Chinese High Technology Letters, 2015, 25(S1):822-828.(in Chinese)
[17] 刘建磊.基于变分模型的图像去雾方法[J].计算机工程, 2017, 43(11):216-220, 226. LIU J L.Image dehazing method based on variational model[J].Computer Engineering, 2017, 43(11):216-220, 226.(in Chinese)
[18] WANG Z Y, YAO Z J, WANG Q Y.Improved scheme of estimating motion blur parameters for image restoration[J].Digital Signal Processing, 2017, 65:11-18.
[19] 雷蕾, 郭东恩, 靳峰.基于谱归一化条件生成对抗网络的图像修复算法[J].计算机工程, 2021, 47(1):230-238. LEI L, GUO D E, JIN F.Image inpainting algorithm based on conditional generative adversarial network with spectral normalization[J].Computer Engineering, 2021, 47(1):230-238.(in Chinese)
[20] 王坤, 刘沛伦, 王力.基于FFWA的自适应Canny飞机蒙皮红外图像边缘检测[J].红外技术, 2021, 43(5):443-454. WANG K, LIU P L, WANG L.Infrared image adaptive Canny edge-detection of aircraft skin based on fast fireworks algorithm[J].Infrared Technology, 2021, 43(5):443-454.(in Chinese)
[21] ADAMSKI M, SARNACKI K, SAEED K.Binary handwriting image enhancement by directional field-guided morphology[J].Information Sciences, 2021, 551:168-183.
[22] PANDIT V R, BHIWANI R J.Morphology-based spatial filtering for efficiency enhancement of remote sensing image fusion[J].Computers & Electrical Engineering, 2021, 89:1-15.
[23] NAH S, KIM T H, LEE K M.Deep multi-scale convolutional neural network for dynamic scene deblurring[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C., USA:IEEE Press, 2017:257-265.

选择文件类型/文献管理软件名称

选择包含的内容