改进SRAD模型在SAR图像斑点噪声滤波中的应用  

Application of Improved SRAD in SAR Image De-Speckling

作　　者：王红卫[1] 郭立群[2,3] 杨洪哲 靳克强[5] 

机构地区：[1]信息工程大学测绘学院 [2]大连理工大学管理学院 [3]65015部队 [4]惠州市市政工程勘察设计研究院 [5]96633部队

出　　处：《测绘科学技术学报》2010年第3期189-192,共4页Journal of Geomatics Science and Technology

摘　　要：斑点噪声的存在严重影响了合成孔径雷达(SAR)图像的应用,所以抑制斑点噪声显得尤其重要。斑点噪声各向异性扩散模型(SRAD)能较好地去除斑点噪声并保持边缘信息,但依然存在不足之处。在分析SRAD模型的基础上,结合增强Lee算法思想,将SAR图像细分为均匀区域、弱纹理区域和噪声发育不完全区域,针对不同区域采用不同处理方法,并加入了迭代终止条件。实验证明改进的SRAD算法具有更好的平滑效果和边缘保持能力。The existence of speckle noise has seriously affected the application of SAR Image,so it is particularly important to reduce speckle.Although Speckle Reducing Anisotropic Diffusion（SRAD） can remove speckle noise as well as preserve edge better,there is still some insufficient sections.On the basis of analyzing SRAD and with the idea of arithmetic of enhanced Lee,the SAR image has been divided in detail into three types of areas： homogeneous area,weak texture area and area with incomplete development noise.Different treatments were used in different types of areas and termination conditions were added into this arithmetic.The experiments have validated that the improved arithmetic have better effect in smoothing and better ability in edge preservation.

关 键 词：合成孔径雷达 斑点噪声 斑点噪声各向异性扩散模型 增强Lee 改进SRAD 

分 类 号：P237.3[天文地球—摄影测量与遥感]