一种自适应阈值曲波图像去噪算法

利用曲波变换进行图像去噪能取得较好的效果,但是现有的曲波去噪没有充分利用图像曲波系数的特点,容易过分扼杀图像的细节信息.本文提出了一种基于阈值自适应的曲波图像去噪算法,利用图像经过曲波变换后的能量分布特性选取阈值,以更好地保护图像细节.实验结果表明,该算法可以将每一尺度上的信号与噪声在最大程度上分离,有效去除了高斯白噪声,进一步提高了峰值信噪比,更好地实现了去除噪声与保护图像细节之间的平衡.

Curvelet Transform Algorithm with Auto-adaptive Threshold to Image Denoising

The curvelet transform is one of the recently developed multiscale transforms.It processes the linear singularity and provides optimally sparse representation of images with edges.But now the image denoising based on the curvelet transform is almost never used by the Monte Carlo threshold.The feature of image curvelet coefficients is not used,so the best result cannot be reached.An algorithm based on auto-adaptive Monte Carlo threshold is proposed.Experimental results on several testing images show that the...