摘要
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。
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.
出处
《微型机与应用》
2010年第3期20-21,25,共3页
Microcomputer & Its Applications
关键词
超分辨率重建
POCS算法
平稳离散小波变换技术
super-resolution reconstruction
projections onto convex sets（POCS） algorithm
stationary discrete wavelet transform
