基于参数估计精度的SAR图像分辨率评估方法  

New Method for Resolution Evaluation of SAR Image Based on Parameter Estimation Precision

作　　者：王卫威[1] 王正明[2] 汪雄良[2] 袁震宇[2] 

机构地区：[1]国防科学技术大学信息系统与管理学院,湖南长沙410073 [2]国防科学技术大学理学院,湖南长沙410073

出　　处：《系统仿真学报》2008年第10期2542-2545,共4页Journal of System Simulation

基　　金：国家自然科学基金(60572136)

摘　　要：分辨率是SAR图像超分辨处理性能评估的一个重要指标。首先分析传统的3dB主瓣宽度计算方法用于超分辨性能评估的缺陷与不足,并利用三次样条方法证明了对传统方法而言,像元分辨率为最高测量分辨率。然后基于频域点散射模型,从最优参数估计精度角度提出了新的分辨率评估方法。该方法克服了传统方法的缺点,并建立了分辨率与目标参数、噪声水平之间的联系,可用于SAR图像超分辨及其它处理性能的预测与评估。利用仿真实验验证了该方法的有效性。The resolution is an important index for evaluating the performance of SAR image super-resolution. The deficiency of the traditional method of 3dB mainlobe width in the application of the super-resolution performance evaluation was analyzed, and it was proved the pixel resolution was the highest measure resolution of this method. Then, a new method for resolution evaluation from the view of optimal performance of parameter estimation was put forward based on the point scattering model in the frequency domain. The new method avoids the deficiency of the traditional one, and establishes the relationship between the resolution and the target parameters, the noise level. It can be used for the performance evaluation of SAR image super-resolution and other processing. A simulation experiment illustrates its validity.

关 键 词：SAR图像 参数估计 克拉美-罗下界 点散射模型 分辨率 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]