基于双树复小波变换与双边滤波的图像滤波

Image denoising based on dual tree complex wavelet transform and bilateral filter

为了进一步提升高斯噪声的去除性能,提出了基于双树复小波变换与双边滤波的图像滤波方法.根据图像和噪声的分布特征,推导出一种自适应的阈值去噪模型.用去噪模型对双树复小波变换后的图像系数进行量化处理,再由双树复小波逆变换得到去噪图像,然后用改进的双边滤波方法对去噪图像进行边缘增强,改进的双边滤波核自适应于图像的特征,具有更好的鲁棒性.实验结果显示,该方法相对于现有的性能较好的方法,PSNR高出大约0.8 dB,SSIM高出大约2.3%.实验证明了该文提出的方法在去噪效果和细节恢复上优于已有的方法.

To break through the bottleneck of the existing Gaussian noise reduction methods,an image denoising method based on dual tree complex wavelet transform and bilateral filter is proposed.Based on the distribution characteristic of the image and noise,an adaptive threshold denoising model is derived.By the adaptive threshold denoising model,the image coefficients by dual tree complex wavelet transform are quantitatively processed,and the denoised image is obtained by the inverse dual tree complex wavelet transform.And then the edge of denoised image is enhanced by the improved bilateral filter,in which the bilateral filtering kernel is adaptive to image features,showing better robustness.The experimental results show that the PSNR achieved by the proposed method is about 0.8 dB higher than that of existing fairly good methods,and the SSIM is about 2.5%higher.The experiments confirm is superior to the existing methods in noise removal and edge restoration.