基于线性回归的SAR目标方位角估计方法

方位角估计是合成孔径雷达(SAR)图像自动目标识别研究中的一个重要问题。一般而言,基于SAR图像的目标方位角估计方法应满足精度高、速度足够快、对目标部署条件的变化具有一定的稳健性。文中在分析现有的SAR图像目标方位角估计方法优缺点的基础上,给出了一种有效的基于线性回归的SAR图像目标方位角估计方法,并通过对大量实测MSTARSAR图像目标方位角的估计试验,详细分析了该方法对SAR图像目标方位角的估计性能。

A Method of Estimating Target Azimuth From SAR Image Based on Linear Regression

Target azimuth estimation is very important for SAR ATR. Generally, any method of target azimuth estimation form SAR image should meet three requirements: First, it should be accurate. second, it should be fast. Finally, it also should be robust against target deployment variation. The advantages and disadvantages of current methods of target azimuth estimation from SAR image are analyzed in this paper. Then a method of estimating target azimuth from SAR image based on linear regression is proposed, and the performance of this estimation method is analyzed through a large number of experiments using real MSTAR SAR target image.