小波自适应比例改进算法在图像去噪中的应用

Image denoising using improved adaptive proportion-shrinking algorithm in Wavelet field

为了改善图像质量和便于识别、压缩等处理,图像去噪是必不可少的过程。提出一种改进的小波比例萎缩去噪算法,摒弃单一根据功率谱构造萎缩函数的方法,在函数式中加入自适应阈值策略。利用局部信号与噪声的小波系数方差比确定阈值,具有局部适应性强,重建误差小的特点。Matlab仿真实验在原始图像加入N(0,202)白噪声,去噪后图像均方误差减小226,峰值信噪比提高了3.722dB;而且重建图像视觉效果更好,并能保持图像边缘细节。

To improve the quality of image and facilitate image further processing - recognition, compression and so on, it is necessary to denoise image. An improved proportion-shrinking algorithm in wavelet denoising was proposed, in which a threshold strategy was added to proportion-shrinking algorithm instead of simplex strategy of signal power. The threshold was educed according to the square error ratio of wavelet coefficient of local signal and noise. It has a good locally adaptive adaptability and a characteristic of less reconstruction error. The N（0, 20^2） Gaussian white noise is added to the original image for MATLAB simulation. Results show the MSE （mean square error） of the denoised image can decrease 226 and the PSNR is improved by 3.722dB. The reconstructed image has better visual effect with clear edge information.