|
|||||
|
|
| 采用离散余弦变换的小波图像去噪方法 | |
| 引用本文： | 芮挺,王金岩,沈春林,丁健.采用离散余弦变换的小波图像去噪方法[J].光电工程,2005,32(1):51-54. |
| 作者姓名： | 芮挺 王金岩 沈春林 丁健 |
| 作者单位： | 1. 南京航空航天大学,自动化学院,江苏,南京,210016;解放军理工大学,工程兵工程学院,江苏,南京,210007 2. 上海交通大学,电子信息与电气工程学院,上海,200030 3. 南京航空航天大学,自动化学院,江苏,南京,210016 4. 解放军理工大学,工程兵工程学院,江苏,南京,210007 |
| 基金项目： | 航空科研项目;空军装备部科研项目 |
| 摘 要： | 提出一种通过对小波域中噪声能量的估计来进行去噪的新方法。算法采用离散余弦变换(DCT)提取小波系数的主要特征,无需对噪声方差进行估计。对图像进行小波分解,利用 DCT对高频子带进行局部特征提取;利用部分 DCT 系数对小波系数进行重建,并以重建系数的平均能量作为局部噪声能量的估计;去除原小波系数中的噪声分量后,进行小波逆变换,得到去噪后的图像。实验证明,其峰值信噪比(PSNR)比通常的阈值萎缩法提高了 2-4dB。 |
| 关 键 词： | 图像处理 小波变换 离散余弦变换 去噪 |
| 文章编号： | 1003-501X(2005)01-0051-04 |
| 收稿时间： | 2004/5/11 |
| 修稿时间： | 2004年5月11日 |
Wavelet image denoising based on discrete cosine transform |
|
| RUI Ting,WANG Jin-yan,SHEN Chun-lin,DING Jian.Wavelet image denoising based on discrete cosine transform[J].Opto-Electronic Engineering,2005,32(1):51-54. | |
| Authors: | RUI Ting WANG Jin-yan SHEN Chun-lin DING Jian |
| Affiliation: | RUI Ting1,WANG Jin-yan3,SHEN Chun-lin1,DING Jian2 2 |
| Abstract: | A new denoising method which extracts feature of wavelet coefficients and then estimates noise energy from this feature is proposed. Firstly, wavelet of noisy image is analyzed, and local feature is extracted with Discrete Cosine Transform (DCT); Secondly, wavelet coefficients are reconstructed based on DCT and local noise energy is estimated based on mean energy of the reconstructed wavelet coefficients; And thirdly, denoising wavelet coefficient is derived though subtracting noise energy from original wavelet coefficient; At last, we can do inverse wavelet transaction to get the denoised image. The algorithm presented in the paper has nothing to do with noise variance and DCT algorithm is very simple. Experimental results demonstrate that peak signal-noise-ratio of this algorithm has been improved for 2-4dB in comparison with common threshold method. |
| Keywords: | Image processing Wavelet Transform Discrete cosine Transform Denoising |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ | |