压缩感知理论在图像去模糊中的应用

Image De-Blurring Based on the Compressive Sensing Theory

把图像去模糊算法与模糊图像质量评价方法相结合,提出一种图像去模糊的算法。该算法在图像去模糊的迭代计算中,设置迭代点,在迭代点处,对去模糊的图像进行质量评价,计算前后两次评价分数之差,若小于给定的期望值,则终止执行图像去模糊操作,算法结束。通过仿真实验,验证了算法的有效性,通过比较迭代点处模糊图像的信噪比PSNR与质量评价分数Score,验证了图像评价分数能够更好地区分图像的视觉质量差异。

In the paper, an algorithm of image de-blurring is proposed by combining the algorithm of the image de-blurring and the method of the blurring image quality assessment. The iteration point is set in the iterative calculation of image de-blurring, at the iteration point, the de-blurred image is assessed, The difference between the two assessing scores is less than a given expected value, then the image deblurring algorithm is terminated, and the algorithm is finished. The effectiveness of this algorithm is verified by simulation experiments. At the iteration point, the power signal-to-noise ratio of the de-blurring image is compared with its assessing score, it is verified that the blurring image assessing score can distinguish the difference of visual quality of images well.