论文部分内容阅读
与小波变换相比,Curvelet变换更好地表达图像的边缘和细节,因此更适合多尺度图像去噪。针对软阈值和硬阈值去噪方法存在的不足,提出了基于Curvelet变换域的软硬阈值折衷去噪法,并采用不同的阈值自适应地对不同的Curvelet子带进行阈值化。实验结果表明该方法对图像中的边缘、弱的直线和曲线特征有更好的恢复。去噪后图像PSNR值更高,视觉效果更好。
Compared with wavelet transform, Curvelet transform to better express the edges and details of the image, so it is more suitable for multi-scale image denoising. Aiming at the shortcomings of soft threshold and hard threshold denoising methods, a soft and hard threshold tradeoff denoising method based on Curvelet transform domain is proposed, and different Curvelet subbands are adaptively thresholded using different thresholds. The experimental results show that this method can recover edge, weak line and curve features in the image better. After de-noising, the PSNR value is higher and the visual effect is better.