Abstract: This paper presents an algorithm of color image denoising based on the second generation Curvelet transform. The wavelet transform has inherent defect in describing directional characteristics of color picture edge. The new algorithm overcomes these defects and is more suitable for analyzing the edge characteristic of curves or straight lines in two-dimensional image. It has better approaching accuracy and spares ability of expression. Results of simulation experiment to 512×512 Lena and Babon color image show the new algorithm is superior to wavelet and Ridgelet algorithm either in visual effect or in performance criteria.

Key words: the second generation Curvelet transform, color image denoising, wavelet transform, Ridgelet transform

摘要： 提出基于第2代Curvelet变换的彩色图像去噪算法，克服小波变换在表达彩色图像边缘的方向特性等方面的内在缺陷。该算法适合分析二维彩色图像中的曲线或直线状边缘特征，且具有较高的逼近精度和稀疏表达能力。通过Matlab对512×512的Lena和Babon彩色图像进行仿真实验，结果表明，该算法在视觉效果和性能指标方面都优于小波和Ridgelet算法。

关键词: 第2代Curvelet变换, 彩色图像去噪, 小波变换, Ridgelet变换

CLC Number: 

• TP391