一种基于NSCT域的自适应阈值函数SAR图像去噪  

A SAR Image Despeckling Using Self-adaptive Threshold FunctionBased on NSCT

作　　者：彭敏[1] 刘文波[1] 张弓[2] 

机构地区：[1]南京航空航天大学自动化学院,江苏南京210016 [2]南京航空航天大学信息科学与技术学院,江苏南京210016

出　　处：《佳木斯大学学报（自然科学版）》2009年第6期815-817,839,共4页Journal of Jiamusi University：Natural Science Edition

基　　金：航空科学基金资助项目(04D52032)

摘　　要：提出一种基于Nonsubsampled Contourlet变换的能量自适应的阈值函数合成孔径雷达图像去噪方法.该算法针对硬阈值函数、软阈值函数的缺点,利用具有冗余性和良好的方向选择性的Nonsubsampled Contourlet变换,结合根据图像不同分解层不同方向的轮廓细节能量的分布自适应调节的阈值函数完成去噪.实验结果表明该算法在有效抑制SAR图像斑点噪声的同时能很好地保持图像的边缘细节特征.A new algorithm which was an energy self-adaptive threshold function for the SAR image denoising was presented in this paper.This algorithm avoided the shortcomings of the hard-threshold function and the soft-threshold function.It combined the shift-invariance with direction selectivity of the nonsubsampled contourlet transform,and also accommodated the threshold function parameter itself by consideration of the image details which have different energy in each subbands and directions.The results of the expeirments on the SAR image denoisnig show that： the proposed algorithm can reduce the speckle noise more effectively while preserving the edges of the SAR image.

关 键 词：合成孔径雷达图像 非抽样轮廓波变换 相干斑抑制 阈值函数 

分 类 号：TN957.52[电子电信—信号与信息处理]