一种改进的毫米波图像超分辨率重建算法

针对毫米波成像时,由于天线孔径受限使得获取的图像空间分辨率很低、高频信息损失严重的问题,提出一种改进的非凸集投影超分辨算法。该算法以非凸集阈值收缩迭代算法(Non-convex Shrinking Iteration,NCSHI)为基础,采用具有平移不变特性的双树复数小波作为稀疏基,引入了两步迭代过程,有效地利用了前两次的迭代信息。实验仿真结果表明,该算法有效地改善了伪吉布斯效应,收敛速度更快,具有良好的超分辨性能。

A Modified Method of Passive Millimeter-Wave Imaging

Abstract: To solve the problem of low resolution images caused by the low-pass effect of passive millimeter wave imaging system, a modified method based on the non-convex shrinking iteration (NCSHI) algorithm for reconstruction process is put forward in this paper. This paper presents an algorithm using dual-tree complex wavelet transform and two-step iterative shrinkage, which makes effective use of the previous estimations in order to gain a more accurate value. Experimental results demonstrate this algorithm can overcome the pseudo-Gibbs effect, has better super-resolution performance and its convergence rate is faster.