基于自适应总变分法的AO图像复原算法研究  

Research on Adaptive Optics Image Restoration Algorithm Based on Adaptive Total Variational Method

作　　者：王爽[1] WANG Shuang(College of Information Technology,Changchun Polytechnic,Changchun 130033,China)

机构地区：[1]长春职业技术学院信息技术分院

出　　处：《红外技术》2019年第12期1106-1110,共5页Infrared Technology

基　　金：教育部职业院校信息化教学指导委员会2018-2020项目（2018LXB0130）;吉林省教育厅科研项目（JJKH20180637KJ）

摘　　要：为了提高自适应光学(AO)图像的质量,研究了一种基于波前重构和自适应总变分(TV)的AO图像复原方法。首先,基于Zernike多项式进行波前重构,对点扩散函数(PSF)进行初始估计。然后,提出了基于自适应总变分法的AO图像复原的迭代求解,解决了联合去卷积问题。最后,通过图像复原实验验证本文算法的恢复效果。实验结果表明:与RL-IBD算法和FS-MLJD算法相比,本文算法的NMSE值分别降低了18.6%、10.7%,算法的PSNR值分别提高了4.47%、0.987%,算法的运算时间分别降低了1.99%、13.66%。In order to improve the image quality of adaptive optics(AO), an AO image restoration algorithm based on wavefront reconstruction and adaptive total variation(TV) is proposed in this work. First, wavefront reconstruction using the Zernike polynomial is employed for an initial estimate of the point spread function(PSF). Then, we developed our proposed iterative solutions based on the adaptive total variation for AO image restoration by addressing the joint deconvolution issue. Image restoration experiments were performed to verify the image restoration effect of our proposed algorithm. Compared with the RL-IBD and FS-MLJD algorithms, the experimental results show the normalized mean square error(NMSE) for a real AO image from our algorithm decreased by 18.6% and 10.7%, respectively, the peak signal-to-noise ratio(PSNR) increased by 4.47% and 0.987%, respectively, and the computation time decreased by 1.99% and 13.66%, respectively.

关 键 词：自适应光学 波前重构 点扩散函数 自适应总变分 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]