基于幸运区域融合的多帧大气湍流图像盲复原  

Blind Restoration of Multi-frame Atmospheric Turbulence Images Based on Lucky Region Fusion

作　　者：鱼轮 王伊君 邢笑笑 YU Lun;WANG Yijun;XING Xiaoxiao(Center for Artificial Intelligence,School of Electronic Information and Electrical Engineering,Shangluo University,Shangluo 726000)

机构地区：[1]商洛学院电子信息与电气工程学院人工智能研究中心,商洛726000

出　　处：《舰船电子工程》2024年第12期39-43,共5页Ship Electronic Engineering

基　　金：陕西省教育厅项目“基于多尺度生成对抗网络的大气湍流退化图像复原研究”(编号:23JK0418);商洛学院科研项目“大气湍流退化图像的复原算法研究”(编号:21SKY010);大学生创新创业训练计划项目“基于稀疏约束的大气湍流退化图像盲复原研究”(编号:S202311396081)资助。

摘　　要：针对大气湍流引起的观测序列图像模糊的问题,提出一种基于幸运区域融合的多帧大气湍流图像盲复原算法。该算法首先在湍流退化序列图像中提取出参考帧,通过光流法去除退化序列中的图像倾斜,然后利用几何一致性和梯度约束进行幸运区域融合,最后采用即插即用型交替方向乘子法交替迭代最小化复原图像。仿真实验结果表明:该算法对图像复原质量有所提升,复原结果具有清晰的边缘轮廓和局部细节。Aiming at the problem of blurred observation sequence image caused by atmospheric turbulence,a blind restora⁃tion algorithm of multi-frame atmospheric turbulence image based on lucky region fusion is proposed.In this algorithm,reference frames are extracted from turbulent degraded sequence images,image skew is removed by optical flow method,geometric consisten⁃cy and gradient constraint are used for lucky region fusion,and the plug-and-play alternating direction multiplier method is used to minimize the restored images.The simulation results show that the algorithm can improve the quality of image restoration,and the restoration results have clear edge contour and local details.

关 键 词：图像盲复原 光流法 幸运区域 交替方向乘子法 

分 类 号：TN911.73[电子电信—通信与信息系统]