Parametric blur estimation for blind restoration of atmospherically degraded images: Class G
- 90 Downloads
Iterative methods are typically utilized for blind image restoration (BIR); however, they are relatively slow, uncertain, and occasionally ill-behaved. This study presents a non-iterative algorithm to estimate the parameters of point spread functions (PSFs), particularly, Class G. We propose a curve model to approximate the normalized spectrum amplitude of the original image in accordance with the decay law of the natural image spectrum. The blur PSF is estimated by comparing the original image spectrum with the degraded one. Then, the image is restored by applying the estimated PSF and the Wiener filter. Experimental results demonstrate that the proposed algorithm can obtain a more accurate PSF and reduce ringing artifacts compared with the existing algorithms. The quality of the restored images is enhanced significantly.
KeywordsBlind image restoration Point spread function estimation Atmospheric turbulence Power law Ringing artifacts
The authors would like to thank Goldstein for sharing his code of the work. This work is funded by the National Natural Science Foundation of China (61303121), the Ministry of Education & China Mobile Joint Research Fund Program (MCM20160302), the Zhongxing Research Grant (3212000210), the National Key Research and Development of China (2016YFB0501300, 2016YFB0501301), and the Major Science and Technology Foundation in Guangdong Province of China (No. 2015B010104002).
- 1.Yitzhaky Y, Dror I, Kopeika N S.: Restoration of atmospherically blurred images according to weather-predicted atmospheric modulation transfer functions. Opt. Eng. 36(11), 3064–3072(1997)Google Scholar
- 3.Chaudhuri, S., Velmurugan, R., Rameshan, R.: Blind deconvolution methods. Springer International Publishing. (2014)Google Scholar
- 4.Vera E, Vega M, Molina R, et al.: Iterative image restoration using nonstationary priors. Appl. Optics, 52(10), D102–D110(2013)Google Scholar
- 17.Shenghua X., Qiheng Z., Ding S.: The improved image restoration algorithm based on APEX method. Laser Infrared, 32(2), 185–188 (2007) (in Chinese).Google Scholar
- 19.Qi G., Hong Z., Kedong W., et al.: Estimation of point spread function for long-exposure atmospheric turbulence-degraded images. Infrared Laser Eng. 43(4), 1327–1331 (2014) (in Chinese).Google Scholar
- 23.Feller, W.: An introduction to probability theory and its applications. John Wiley & Sons, (2008)Google Scholar
- 24.Shi Y., Hong H., Song J., et al.: Blind image deblurring with edge enhancing total variation regularization. Int. Soc. Optics Photonics, 95222B, 7 pages (2015)Google Scholar
- 25.Balasubramanian, M., Iyengar, S. S., Reynaud, J., et al.: A ringing metric to evaluate the quality of images restored using iterative deconvolution algorithms. Proceeding of the 18 International Conference on Systems Engineering (IEEE ICSEng 05), USA: 483–488 (2005)Google Scholar
- 27.Goldstein, A., Fattal, R.: Blur-kernel estimation from spectral irregularities[C]. European Conference on Computer Vision. Springer Berlin Heidelberg, 2012: 622–635Google Scholar