基于假设检验的SAR图像分辨率判别准则与算法

New Criteria and Arithmetic for Calculating the Resolution in SAR Images Based on Hypothesis Test

SAR图像分辨率是衡量SAR系统成像质量的重要指标。常用的3dB主瓣宽度测量方法来源于光学成像中的瑞利准则,受人类视觉感知能力限制而且没有考虑噪声影响。本文基于假设检验理论,根据点目标散射模型建立相位历史域、复数据域以及图像域的两点目标二维分辨模型,得到统计意义上分辨率的计算表达式,并由此得出分辨率与图像质量或噪声水平之间的定量关系。该表达式与现有的分辨率定性理解相一致,克服了传统分辨率定量评价准则的缺陷。

SAR Image Resolution is a very important index for evaluating the imaging quality of SAR system. It is usually gained by calculating 3dB mainlobe width of the point target impulse response. This method derives from the Rayleigh Criteria,which is limited by the ability of the human vision and doesn＇ t deal with the effect of the noise. Based on the hypothesis test theory, this paper establishes the resolution criteria and models of two point targets in the phase history domain, complex data domain and image domain. As a result, it gets the expression of SAR image resolution in statistical sense, which shows the quantitative relationship between the resolution and the image quality or the noise. The calculating results are consistent with the actual qualitative comprehension about the SAR resolution and this new method overcomes the defects of the traditional evaluating criteria.