匀速直线运动模糊图像的模糊参数鲁棒识别

赵鹏1; 曹军2; 韦兴竹1

doi:10.3788/OPE.20132109.2430

您当前的位置：

首页 >

文章列表页 >

匀速直线运动模糊图像的模糊参数鲁棒识别

信息科学 | 更新时间：2020-08-12

- 匀速直线运动模糊图像的模糊参数鲁棒识别
- Identification of robust blur parameter for uniform linear motion blurred images
- 光学精密工程 2013年21卷第9期页码：2430-2438
- 作者机构：
  
  1. 东北林业大学信息与计算机工程学院，黑龙江哈尔滨 150040
  2. 东北林业大学机械工程博士后流动站，黑龙江哈尔滨 150040
- 作者简介：
- 基金信息：
  
  教育部中央高校基本科研业务费专项基金();科技部973计划立项();教育部新世纪优秀人才支持计划专项();中国博士后科学基金第5批特别资助()
- DOI：10.3788/OPE.20132109.2430
  中图分类号： TP391.4
- 收稿日期：2013-03-31，
  
  修回日期：2013-04-23，
  
  网络出版日期：2013-09-30，
  
  纸质出版日期：2013-09-15
- 稿件说明：
移动端阅览
赵鹏曹军韦兴竹. 匀速直线运动模糊图像的模糊参数鲁棒识别[J]. 光学精密工程, 2013,21(9): 2430-2438

ZHAO Peng CAO Jun WEI Xing-zhu. Identification of robust blur parameter for uniform linear motion blurred images[J]. Editorial Office of Optics and Precision Engineering, 2013,21(9): 2430-2438
赵鹏曹军韦兴竹. 匀速直线运动模糊图像的模糊参数鲁棒识别[J]. 光学精密工程, 2013,21(9): 2430-2438 DOI： 10.3788/OPE.20132109.2430.

ZHAO Peng CAO Jun WEI Xing-zhu. Identification of robust blur parameter for uniform linear motion blurred images[J]. Editorial Office of Optics and Precision Engineering, 2013,21(9): 2430-2438 DOI： 10.3788/OPE.20132109.2430.

摘要

针对摄像机成像时经常产生匀速直线运动模糊，导致图像退化的现象，提出了恢复图像涉及的精确辨识运动模糊尺度的方法。由于利用Radon变换可以精确地识别运动模糊方向，通过图像旋转可以将运动模糊的方向旋转到水平轴，因此，只对水平方向的运动模糊图像进行研究。对于不带噪声的运动模糊图像，对其进行Fourier变换转化到频率域，使用BP神经网络检测运动模糊尺度；BP神经网络的输入量为频谱图中央区域的幅度加和。对于带噪声的运动模糊图像，先对其进行双谱变换，再使用BP神经网络检测运动模糊尺度；双谱中每列的最大值为BP神经网络的输入量。最后通过仿真实验验证了本文方法的正确性和有效性。验证结果显示，当噪声图像的信噪比SNR23 dB时，本文方法的模糊参数辨识平均误差5%，优于传统的运动模糊参数辨识方法。

Abstract

Uniform linear motion blur often occurs in the camera imaging，which results in image degradation seriously. Therefore

this paper proposes a method to identify the motion blur extent accurately. Since a motion blur angle could be identified accurately by Radon transform and the motion blur direction could be rotated to the horizontal

only the horizontal motion blurred images are needed to be researched. For a motion blurred image without noises

the Fourier transform is used to transform it to frequency domain

the motion blur length is estimated by the BP neural network and the input of BP network is the sum of the amplitudes in the central region of the spectrum. For a blurred image with noises

the bispectrum transform is applied

the motion blur length is estimated by BP neural network and the input of BP network is the maximum value of each column of bispectrum. The simulation experiments show that the proposed method in this paper is correct and efficient. When the SNR is larger or equal to 23 dB in noisy images

the mean error of blur parameter identification by proposed method is below 5%

which is superior to that of conventional schemes.

关键词

Keywords

references

梁炯，许廷发，石明珠,等.一种提高匀速直线运动参数辨识精度的方法[J].北京理工大学学报，2012, 32（8）： 818-823.LIANG J, XU T F, SHI M ZH, et al.. A method to improve accuracy of parameter identification for single image blurred by uniform linear motion [J]. Transactions of Beijing Institute of Technology, 2012, 32(8): 818-823. (in Chinese)[2]JUNSEONG P, MIN K, SOONKEUN C, et al.. Estimation of motion blur parameters using cepstrum analysis [C]. 2011 IEEE 15th International Symposium on Consumer Electronics(ISCE), 2011, 406-409. [3]HE W G, LI S F, HU G W. Blur identification using an adaptive adaline network [C]. Proceedings of 2005 International Conference on Machine Learning and Cybernetics, 2005,9: 5314-5317. [4]XU T F, PENG Z. Objects translational speed measurement using motion blur information[J]. Measurement, 2010, 43(9)： 1173-1179. [5]郭永彩，郭瑞瑞，高潮.运动模糊图像点扩展函数的参数鉴别[J]．仪器仪表学报，2010，31（5）：1052-1057．GUO Y C, GUO R R, GAO C. Identification of blur parameters from motion blurred images [J]. Chinese Journal of Scientific Instrument, 2010，31(5)：1052-1057．(in Chinese)[6]李楠，路小波.散焦和运动模糊复合模型及参数估计[J]．仪器仪表学报，2011，32（11）：2523-2529．LI N, LU X B. Hybrid model and parameter estimation combining defocus and motion blur [J]. Chinese Journal of Scientific Instrument, 2011，32(11)：2523-2529． (in Chinese)[7]LI Q, YOSHIDA Y. Parameter estimation and restoration for motion blurred images [J]. IEICE The Institutes of Electronics, Information and Communication Engineers, 1997, E80A(8):1430-1437. [8]CANNON M. Blind deconvolution of spatially invariant image blurs with phase [J]. IEEE Transactions on Acoustics, Speech and Signal Processing, 1976, 24(1), 58-63. [9]BHASKAR R, HITE J, PITTS D E. An iterative frequency-domain technique to reduce image degradation caused by lens defocus and linear motion blur [C]. Proc. of IEEE IGARSS, 1994,4:2522-2524. [10]CHANG M M, TEKALP A M, ERDEM A T. Blur identification using the bispectrum [J]. IEEE Trans. on Signal Processing，1991，39(10)：2323-2325．[11]MAYNTX C, AACH T, KUNZ D. Blur identification using a spectral inertial tensor and spectral zeros [C]. 1999 International Conference on Image Processing, 1999,2:885-889. [12]MOGHADDAM M E, JAMZAD M. Motion blur identification in noisy images using mathematical models and statistical measures [J]. Pattern Recognition, 2007, 40(7), 1946-1957.[13]MOGHADDAM M E, JAMZAD M. Linear motion blur identification in noisy images using fuzzy sets [J]. EURASIP Journal on Adv. Signal Process, 2007, doi:10.1155/2007/68985. [14]TOFT P. The Radon transform-theory and implementation [D]. Denmark: Technical University，1996．[15]马锐．人工神经网络原理[M]．北京：机械工业出版社，2010．MA R. Artificial Neural Network Theory [M]. Beijing: China Machine Press, 2010. (in Chinese)[16]HUI J, CHAO Q L. Motion blur identification from image gradients [C].Proc. of IEEE CVPR, 2008, Vol.1, 1-8.

浏览量

179

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

中红外波段超构透镜研究进展

特征融合和自适应权重更新相结合的运动模糊目标跟踪

基于增强拉东变换及稀疏先验的模糊星斑复原

高动态环境下的运动模糊图像配准

单一原子氧辐照对SR107-ZK白漆反射率的影响