频域滤波技术在图像复原中的应用研究

ON APPLYING FREQUENCY DOMAIN FILTERING TECHNOLOGY TO IMAGE RESTORATION

对不同的频域滤波图像复原技术的图像复原效果进行研究,对现有Lucy_Richardson频域滤波图像复原算法存在的问题进行改进,提出图像降噪和图像复原迭代交替进行的图像复原方法。分别对LR算法和改进后PLR算进行仿真实验,通过实验证明改进后的PLR算法不仅提高了图像复原运算的速度,而且提高了图像恢复的质量。同时还结合PSNR图像质量客观评价法和HVS对比度评价法对复原图像质量进行客观评价,两种方法结合使用克服了传统上采用PSNR评价方法造成评价结果和人眼观察结果不一致的缺陷。评价结果均表明改进后的PLR算法在迭代次数相同的条件下图像恢复质量均优于LR算法,并且可以消除复原图像产生振铃现象的发生。

We study the effects of image restoration using the technology of image restoration with different frequency domain filtering,make the improvement in regard to the problems of existing Lucy_Richardson frequency domain filtering image restoration algorithm,and propose a new image restoration method in which the image noise reduction and the image restoration are alternatively iterated.Then we implement the simulation experiments on the LR algorithm and the improved PLR algorithm respectively,through the experiments we verify that the improved PLR algorithm does not only improve the speed of image restoration operation,but also improves the quality of the image restoration.In this paper we objectively evaluate the quality of the restored images by combining PSNR objective evaluation method with HVS contrast method,the integrated use of the two methods overcomes the defects of inconsistency between the results derived from traditionally used PSNR evaluation method and the results from human eyes observation.Evaluation results indicate that the improved PLR algorithm outperforms the LR algorithm in image restoration qualities under the condition of same iteration times,and can eliminate the ringing effect caused by image restoration.