基于粒子群优化算法的盲图像复原技术

On Blind Image Restoration Technique Based on Particle Swarm Optimization

针对高斯模糊及失焦模糊图像,提出利用粒子群优化算法鉴别模糊图像的PSF,再将鉴别后的PSF与模糊图像进行Wiener滤波复原,得到估测复原图像,并计算其目标函数值,判断图像是否清晰,决定粒子的演化方向。根据PSO的演化机制,经过N个迭代的计算,粒子最后会收敛在最佳解上,此时得到的估测复原图像最接近原始图像。仿真实验表明,本算法比其他复原算法具有更好的复原效果。

Based on the Gauss blur and Out-of-focus blur image, this article proposes the use of particle swarm optimization to identify fuzzy image of the PSF, and then restores with identified PSF and fuzzy image by using Wiener filter, gets the estimated image, and calculates the value of the objective function,determines whether the image is clear and the direction of particle evolution. According to the evolution mechanism of PSO, by the calculation of N iteration, particle will finally converge in the optimal solution.The experiment shows that the obtained estimation of image restoration is closest to the original image.