合成孔径雷达目标图像复原方法仿真研究

摘要

对于孔径雷达目标图像的复原,可以有效提升合成孔径雷达对地物目标信息的获取能力.对雷达目标图像进行复原,需要通过控制目标图像复原参数,得到精确的合成孔径雷达目标图像模糊核,完成对目标图像的复原.传统方法先利用图像显著边缘扩散函数修正图像位置,抑制图像波动,但无法得到目标图像的模糊核,导致复原效果不理想.提出基于半二次规整化理论的合成孔径雷达目标图像复原方法.利用双边滤波器对合成孔径雷达目标图像进行去噪,对去噪滤波处理后的图像高强度反差边缘,当前尺寸的模糊核作为为初始值,依据尺度大小控制目标图像复原参数,得到精确的合成孔径雷达目标图像模糊核,对其进行模糊核复原迭代求解,由此完成对合成孔径雷达目标图像复原.实验结果表明,所提方法能够有效降低噪声水平,在合成孔径雷达目标图像复原效率方面有很大提高.%Traditional image restoration method uses function of salient edge spread to amend image position.It cannot obtain fuzzy kernel of target image,which leads to unsatisfactory effect of restoration.Therefore,a restoration method for target image of synthetic aperture radar is proposed based on theory of half-quadratic regularization.The two-sided filter is used to de-noise the target image,then the edge of image with high intensity after de-noising and filtering is contracted.The restoration parameter of target image is controlled according to scale by using fuzzy kernel with current size as initial value,and precise fuzzy kernel is obtained.Finally,the fuzzy kernel is solved by using restoration iteration.Thus,the restoration of target image is completed.Simulation results show that the method can reduce noise level effectively.It improves restoration efficiency apparently.