基于POCS算法的图像超分辨率重建

传统POCS算法对图像进行超分辨率重建时,一般都假设所处理的噪声为零均值的加性高斯白噪声,当噪声为非高斯噪声如椒盐噪声时,POCS算法的重建效果将会下降.针对这一问题,本文对含噪图像首先采用平稳离散小波变换技术进行去噪预处理,然后再用POCS算法重建图像.实验证明,此方法对信噪比较低的图像有很好的重建效果,对高斯及椒盐等噪声处理比较有效.

Reconstruction of image super-resolution based on POCS algorithm

Traditional POCS algorithm in super-resolution image reconstruction is generally handled by assuming the noise as zero mean additive white Gaussian noise.When the noise is non-Gaussian noise,for example,the salt & pepper noise,the reconstruction effect of POCS algorithm will decline.To solve this problem,by the noisy image containing a smooth discrete wavelet transform de-noising pre-processing techniques,POCS algorithm is applied for image reconstruction.Experiments show that the method proposed in this paper for low SNR images has a very good reconstruction results,which is more effective than Gaussian and salt & pepper noise processing.