针对当前图像降噪技术难以保留边缘平滑区和高频变化区的纹理细节的不足,提出联合双边滤波器(joint bilate-ral filter,JBF)与多分辨率分析的图像降噪算法.利用参考图像作为参数估计,根据参考图像的丰富高频信息,考虑像素间的距离与亮度相似性,准确计算联合双边滤波器的权值,有效避免在滤波过程中对边缘平滑;引入小波变换、轮廓波变换和非下采样轮廓波变换的多分辨率分析(MRA),解决不同分辨率下的不同频率的图像降噪.实验结果表明,与当前图像降噪技术相比,该算法拥有更好的视觉效果,能较好地保留平滑区和高频变化区的细节信息.%To solve the problem of preserving the edge and the texture details of the high frequency region in the Gauss noise re-moval,the joint bilateral filter (JBF)and a multi-resolution image denoising method were proposed.A reference image was used as parameter estimation,based on the rich high frequency information of the reference image,the distance between pixels and the luminance similarity was considered,and the weight of the joint bilateral filter was calculated.Therefore,the smoothing of edges was avoided.The multi-resolution analysis consisting of the wavelet transform,the contour transformation and non-subsampled contourlet transform (MRA)was introduced,which solved different resolutions of image denoising.Experimental results show that the proposed algorithm has better visual effects,and can better retain the details of the smooth area and high frequency change area compared with the current image denoising technology.
展开▼