基于Contourlet域SOT结构的SAR图像相干斑抑制算法  

Speckle Denoising Algorithm of SAR Image Based on SOT Structure in Contourlet Domain

作　　者：张绘[1] 张弓[1] 郭琦南[1] 

机构地区：[1]南京航空航天大学信息科学与技术学院,南京210016

出　　处：《南京航空航天大学学报》2006年第6期743-748,共6页Journal of Nanjing University of Aeronautics & Astronautics

基　　金：江苏省自然科学基金(BK2001047)资助项目;航空科学基金(04D52032)资助项目

摘　　要：把SOT结构引入到能够最优表示图像的Contourlet变换中,结合Contourlet系数域的结构特点和SAR图像相干斑乘性噪声模型提出了一种新的相干斑抑制算法。对基于小波系数域SOT结构的相干斑抑制算法中固定在一个低频尺度中选择根系数的方法进行了改进,根据SAR图像的匀质性条件自适应地在两个不同低频尺度中选择根系数。把该算法应用到实测SAR图像中取得了很好的目视效果,各项指标得到了有效提高。An algorithm for reducing speckle noise on the synthetic aperture radar （SAR） image is pre sented based on the spatial oriented tree （SOT） structure in a Contourlet domain. The SOT structure is introduced to the Contourlet transform which can represent images. Combining the characteristics of the Contourlet domain and the muhiplicative model of the speckle noise on SAR images, the new denoising algorithm is proposed. The algorithm improves the denoising method based on the SOT structure in the wavelet domain and root coefficients are chosen in a fixed low frequency scale. The improved method chooses root coefficients in two different low frequency scales adaptively according to the homogeneous condition on SAR images. By applying the algorithm to the real SAR images, the denoised image obtains a pretty well visual effect and all indices are enhanced.

关 键 词：图像处理 相干斑抑制 CONTOURLET变换 SOT结构 SAR 

分 类 号：TP751.1[自动化与计算机技术—检测技术与自动化装置]