基于L0梯度范数泛函优化和变换域阈值的图像去噪算法  

Image Denoising Algorithm Based on L0 Gradient Minimization and Threshold Shrinkage

作　　者：赵秀芝[1] 

机构地区：[1]浙江工贸职业技术学院电子工程系,浙江温州325003

出　　处：《浙江工贸职业技术学院学报》2014年第1期36-39,共4页Journal of Zhejiang Industry & Trade Vocational College

基　　金：2013浙江省高等教育课堂教学改革研究项目(kg2013836)

摘　　要：基于L0梯度泛函优化和变换域阈值法的图像去噪算法,首先利用L0梯度泛函的最优化算法把含有噪声的图像分解为显著边缘层和细节纹理细节层,然后对含有噪声的细节纹理层图像进行短时傅利叶转换,并在变换域中利用阈值法分离纹理细节层中的图像细节纹理和噪声,进而去除图像的噪声。实验结果表明图像去噪算法获得了显著的去噪效果。An image denoising algorithm based on L0 gradient minimization and threshold shrinkage was proposed in this paper. In this algorithm, the L0 gradient minimization algorithm was firstly used to decompose the noisy image into the large-scale edge lay-er and the fine-scale texture layer and then noises in the fine-scale texture layer was denoised by a threshold shrinkage algorithm on the short-time fourier transform space. The experiment results show that the proposed image denoising algorithm achieves significant improvement in its performance.

关 键 词：去噪 变换阈值 图像分解 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]