摘要
针对远距离成像系统容易受到大气湍流中压强、温度等因素的影响,得到的图像序列会产生图像畸变与模糊等问题,提出了一种基于物理光流与引导图滤波器细节增强的方法,从受湍流影响图像序列中恢复出清晰图像。首先,利用物理光流计算序列帧与参考图像之间的位移信息差,通过位移校正获得去扭曲的图像;其次,采用归一化稀疏表示进行去模糊;最后,利用引导图滤波器将细节增强图像从去扭曲图像中提取出来。实验结果表明,该方法能有效去除湍流图像的畸变和模糊,获得相对清晰的图像,并与其他方法进行对比,证明了其优越性。
In the long-distance imaging system,image sequence was easily affected by atmospheric turbulence.In order to restore images from the turbulence degraded sequence,a method based on optical flow method and guide-image filter was proposed.Firstly,the optical flow method was used to calculate the motion difference between the sequence frames.Image interpolation processing was conducted to obtain a non-distortion frame,by using the motion information.Secondly,sparse representation and guided filters was used to calculate the final de-blurring image from the single non-distortion image.Experimental results showed that the proposed method could reduce the effects of atmospheric turbulence effectively,including temperal-spatial distortion and random blurring.The superiority of this method was proved through the comparison with other methods.
作者
刘亮
蔡泽民
赖剑煌
LIU Liang;CAI Zemin;LAI Jianhuang(Department of Electronic Engineering, Shantou University, Shantou 515063, China;School of Data Science and Computer Science, Sun Yat-Sen University, Guangzhou 510275, China)
出处
《郑州大学学报（理学版）》
CAS
北大核心
2021年第1期47-53,共7页
Journal of Zhengzhou University:Natural Science Edition
基金
国家自然科学基金项目(61876104)
广东省自然科学基金项目(2018A030307034)。
关键词
去扭曲
物理光流
稀疏表示
盲反卷积
引导图滤波器
atmospheric turbulence
optical flow
sparse representation
blind deconvolution
guide image filter
