综合孔径辐射计偏微分方程近场图像反演算法

Near field image reconstruction algorithm of synthetic aperture radiometer based on partial differential equations

针对综合孔径辐射计傅里叶变换成像算法在近场条件下失效的问题,提出了一种基于局部自适应偏微分方程的精确数值图像反演算法.该反演算法根据待反演亮温图像的局部特征,在图像背景区域进行各向同性扩散来抑制噪声,在图像目标区域进行自适应扩散来保持图像边缘细节.近场仿真结果表明,该算法有效地降低了可视度测量噪声对亮温图像反演的影响.用一套8 mm波段二维综合孔径辐射计BHU-2D-U进行了成像实验,实验结果证明了该近场图像反演算法的有效性.

Since the Fourier imaging theory of the synthetic aperture interferometric radiometer（ SAIR） is unavailable under near field conditions,a nontrivial numeric image reconstruction algorithm based on the local adaptive partial differential equations（ PDE） was proposed. Based on the prior information of the desired brightness temperature distribution,an isotropic diffusion was applied at the image background areas to denoise,whereas adaptive diffusion was performed at image target areas to maintain image edges. It was demonstrated by near-field simulation that this algorithm is effective in lowering the image reconstruction errors caused by visibility measurement noise. Imaging experiments were conducted on the two-dimensional 8 mm band SAIR BHU-2D-U,and the advantages of this method were proved.