经过七十多年的发展，合成孔径雷达(Synthetic Aperture Radar ,SAR)因为其独特的信息获取能力，在战场侦察、地形探测等方面发挥着重要的作用。而对SAR的干扰在现代电子对抗(Electronic Countermeasure,ECM)中已经越来越成为学者们关注的热点问题。SAR的欺骗式干扰缺乏与电磁计算的有效结合，存在干扰背景与真实地形背景相似度较低，干扰背景单一，干扰停留在点目标上，干扰效果缺乏有效且贴合实时的评估方法等问题。本文将小斜率近似法应用于对SAR的干扰中，目的在于实现一种基于真实场景的精细化干扰。主要的工作内容如下：

1.为解决干扰背景与真实地形相似度较低的问题，从回波统计和SAR成像的角度，研究真实场景的电磁特性。分析尺度参数和形状参数对杂波统计模型概率分布曲线和回波幅度的影响。提取真实场景的数字高程数据，将其划分为无数个小面元，利用小斜率近似法计算电磁散射系数，并从电磁后向散射系数和不同地形在SAR成像中的亮暗分布与统计模型 进行对比。最后，基于小斜率近似法对真实的沙漠、草地、山地的回波进行计算，拟合对应的回波统计模型参数。结果表明，小斜率近似法计算得到的回波具有统计性符合雷达回波的统计模型，并能够在SAR成像中表现出不同地形的特征，与真实的地形成像相似度较高。

2.基于不同场景的电磁特性对SAR实现精细化的欺骗式干扰。建立SAR欺骗式干扰的几何模型，分析电磁散射分布增量法的基本原理，并从点目标进行仿真说明增量法的有效性。利用小斜率近似法计算三种地形的电磁后向散射系数，并对三种地形进行成像，分析不同地形在SAR成像中亮暗分布的变化。然后在沙漠、草地、山地地形上通过增量法调制散射系数实现坦克、飞机等假目标干扰。最后，将散射分布增量法改进到灰度域，将沙漠和草地的SAR成像图用于对真实SAR图像的干扰，并且在真实的机场图像上调制了假目标干扰。解决了干扰背景单一、干扰停留在点目标上的问题，提高了干扰场景的相似度和精细度。

3.对基于电磁特性的精细化干扰进行评估。研究图像的灰度特征和纹理特征，并从灰度和纹理对精细化的干扰进行评估，对比干扰前后图像灰度的均值、方差、三阶矩和纹理的对比度、能量、熵等参数。结果表明，干扰前后数值变化在允许范围内，验证了干扰图像与原图像的高相似度和干扰的精细化。并利用卷积神经网络对真实SAR图像的场景欺骗式干扰进行智能化的评估。结果表明，智能系统对基于电磁特性的精细化干扰识别率较高。

After more than 70 years of development, Synthetic Aperture Radar (SAR) plays an important role in battlefield reconnaissance and terrain detection because of its unique information acquisition ability. For SAR jamming in modern Electronic Countermeasurehas become the hot issue of scholars attention. The deception jamming of SAR lacks effective combination with electromagnetic calculation, there are some problems, such as low similarity between jamming background and real terrain background, single jamming background, jamming stays on point target, and lack of effective and real-time evaluation methods for jamming effect. In this thesis, the Small Slope Approximation method is applied to SAR jamming, aiming at achieving a fine jamming based on real scene. The main work contents are as follows:

1. In order to solve the problem of low similarity between interference background and real terrain, electromagnetic characteristics of real scene are studied from the perspective of echo statistics and SAR imaging. The effects of scale parameters and shape parameters on probability distribution curve and echo amplitude of clutter statistical model are analyzed. The digital elevation data of the real scene is extracted and divided into countless small surface elements. The electromagnetic scattering coefficient is calculated by using the Small Slope Approximation method, and the electromagnetic backscattering coefficient and the light and dark distribution of different terrain in SAR imaging are compared with the statistical model. Finally, the echo of real desert, grassland and mountain is calculated based on the small slope approximation method, and the corresponding statistical model parameters are fitted. The results show that the echo calculated by the small slope approximation method is statistically consistent with the statistical model of radar echo, and can show different terrain features in SAR imaging, and has a high similarity with the real terrain imaging.

2. SAR can achieve refined spoofing based on electromagnetic characteristics of different scenes. The geometric model of SAR spoofing interference is established, the basic principle of electromagnetic scattering distribution increment method is analyzed, and the effectiveness of increment method is demonstrated by simulation from point target. The electromagnetic backscattering coefficients of three landforms were calculated by using the Small Slope Approximation method, and the three landforms were imaged to analyze the changes of light and dark distribution of different landforms in SAR imaging. Then the scattering coefficient is modulated by increment method to realize the jamming of tank and aircraft on desert, grassland and mountain terrain. Finally, the scattering distribution increment method is improved to the gray domain, and the SAR images of desert and grassland are used to interfere with the real SAR images, and the fake target interference is modulated on the real airport images.The problem that the interference background is single and the interference stays on the point target is solved, and the similarity and fineness of the interference scene are improved.

3. Evaluate refined interference based on electromagnetic characteristics. The gray feature and texture feature of image are studied, and the interference of gray level and texture on refinement is evaluated, and the mean, variance, third-order moment of gray level of image and the contrast, energy and entropy of texture before and after interference are compared. The results show that the numerical variation before and after jamming is within the allowable range, which verifies the high similarity between the jamming image and the original image and the precision of the jamming. The convolution neural network is used to evaluate the scene spoofing of real SAR images intelligently. The results show that the intelligent system has a high recognition rate for the refined interference based on electromagnetic characteristics.