基于边缘检测的NSCT自适应阈值图像去噪

Image Denoising with AT-NSCT Based on Edge Detection

为在去噪时能较好保留图像边缘特征,并针对Coutourlet变换缺乏平移不变性和传统阈值法的不足,提出了一种基于边缘检测的非子采样Contourlet变换自适应阈值(AT-NSCT)图像去噪方法。结果不仅能消除因Contourlet变换缺乏平移不变性而产生的图像失真,而且有效地保留了图像的边缘信息,提高了去噪后图像峰值信噪比,视觉效果更好。

Due to the lack of translation invariance of contourlet transform and traditional threshold＇s disadvantage, and in order to retain the characteristics of the image edge while image denoised, an method of image denoising （AT-NSCT） based on edge detection was proposed. The result not only can eliminate the image distortion because of the lack of translation invariance of Contourlet trans- form, but also maintain the edge information of the image availably. The PSNR of denoised image was improved. And image have better visual effects.