基于随机阵列的NUFFT BP三维快速成像

传统的二维实阵列用于三维微波成像时,需要相对较高的脉冲重复频率才能切换发射阵元。针对这种情况,利用随机阵列来实现三维成像,目的是有效降低脉冲重复频率。进一步提出利用等效相位中心原理来实现任意随机阵列布局的可行性方案。然后,基于随机阵列,发现使用非均匀快速傅里叶变换(nonuniform fast Fourier transform,NUFFT)来实现后向投影(back projection,BP)算法的插值过程,能够用较少的运算量实现BP算法的高精度插值。最后,结合并行处理技术来实现NUFFT BP算法,结果使得BP算法的执行效率得到显著提高。

Fast 3D imaging method for random array based on NUFFT BP

Traditional two dimensional (2D) antenna array requires high pulse repetition frequency (PRF) to switch transmitters, when it is applied to 3D microwave imaging. Therefore, 3D imaging is gotten by the random antenna array, as a result it reduces the requirement for PRF. Besides, a reasonable plan based on the phase center approximation (PCA) is proposed to implement the random antenna array of arbitrary geometry. The back projection (BP) algorithm interpolated by the nonuniform fast Fourier transform (NUFFT) is called NUFFT BP, and achieves high precision interpolation with relatively less computational complexity. Moreover, the NUFFT BP is further accelerated by the compute unified device architecture (CUDA) parallel computing technology, named CUDA NUFFT BP. It significantly improves the execution efficiency of the BP algorithm.