摘要合成孔径雷达(Synthetic Aperture Radar, SAR)是一种微波传感器，由于SAR是一种主动式传感器，它生成SAR图像的原理是由它向地表不断的发射脉冲信号，然后接收从地表反射回的信号。SAR图像相干斑噪声的产生正是由于它特殊的成像机制而导致的，这使得生成的SAR图像信噪比下降，为后期的目标识别和特征提取研究造成了相当大的困难，因此去除相干斑噪声在图像的后期处理中是一个不可缺少的过程。本文主要对去噪后图像的平均后向散射系数,标准差,有效视数,边缘指数这四个评价指标进行对比，另外，就去除SAR图像相干斑噪声，本文主要对LEE与增强型LEE滤波，FROST与增强型FROST滤波，GAMMA滤波,KUAN滤波,SIGMA滤波这七种方法进行研究分析，具体研究结论如下：76482

（1）通过实验数据的对比，有效视数值最大的三种滤波方法为增强型LEE滤波、增强型FROST滤波和GAMMA滤波，同时三种方法的平均后向散射系数值都较为接近原图像；

（2）标准差最小的是GAMMA滤波。边缘保持指数值最大的是SIGMA滤波，即其图像的边缘保持能力最强，但它的标准差较大且有效视数值较小，说明GAMMA滤波的图像去噪能力相对较弱；

（3）与增强型FROST滤波和GAMMA滤波相比，增强型LEE滤波的边缘保持能力相对较强。因此增强型LEE滤波的综合性能是最好的。

毕业论文关键词：SAR图像滤波，滤波方法，滤波方法对比

Abstract Synthetic Aperture Radar (SAR) is a microwave sensor, because SAR is an active sensor, it generates the principle of SAR image is to the surface is continuously transmit pulse signal, and then to receive signals reflected from the surface to the。 SAR image speckle noise is due to its special imaging mechanism, which makes the generation of SAR image signal-to-noise ratio decreased, target recognition and feature extraction in the later caused considerable difficulties, so the removal of speckle noise in the latter part of the image processing is an indispensable process。 This paper mainly to the denoised image of the average scattering coefficient and the standard deviation, the effective number of looks, the edge of the index of the four evaluation indicators were compared, in addition, removal of the SAR image speckle noise, this paper mainly of Lee and enhanced Li Lvbo, the frost with enhanced frost filter, gamma filter wave, wide filter, sigma filter these seven methods of research and analysis, specific conclusions are as follows:

(1) through the comparison of the experimental data, the effective values of maximum the denoising effects of the best three filtering method is enhanced Li Lvbo, enhanced frost filter and gamma filter, also after a mean of three kinds of method to scattering values are relatively close to the original image

(2) the smallest standard deviation is the gamma filter。 The maximum number of edges is the sigma filter, that is, the edge of the image has the strongest ability, but its standard deviation is larger and the effective value is smaller, which indicates that the image denoising ability is relatively weak。

(3) compared with the enhanced frost filter and the gamma filter, the enhanced Li Lvbo's edge preserving ability is relatively strong and therefore enhanced Li Lvbo's comprehensive performance is the best。

Key words: SAR image filtering, filtering method, filtering contrast

目录

摘要 I

Abstract - 1 -

1。1选题背景与意义 - 2 -

1。2SAR图像斑点滤波的研究现状 - 3 -