摘要： 提出了一种基于Curvelet 变换来抑制合成孔径雷达(SAR)图像中噪声的方法，Curvelet 是一种新的多尺度变换理论，具有各向异性的特征，克服了小波在处理大于一维的高维信号时的不足。该文在介绍Curvelet 变换理论及其实现的基础上，引出了SAR 图像斑点噪声的去除方法，并改进了Curvelet 算法，降低了其运算复杂度，讨论了SAR 图像噪声方差的估计，最后和其它的SAR 图像去噪方法作了对比和分析。

关键词: Curvelet 变换；Ridgelet 变换；小波；斑点噪声；去噪

Abstract: A denoising method of SAR image based on Curvelet transform is proposed in this article. Curvelet is a new multiscale transform theory. Curvelet overcomes the limitation of wavelet in analyzing signals with dimension higher than 1-D because it has the character of anisotropy.This article introduces Curvelet theory first and then derivates speckle reducing method of SAR image. It improves the algorithm of Curvelet to reduce its complexity. Comparison is also made with other filtering methods.

Key words: Curvelet transform; Ridgelet transform; Wavelet; Speckle noise; Denoising