«上一篇
 文章快速检索 高级检索

 雷达学报  2018, Vol. 7 Issue (3): 376-386  DOI: 10.12000/JR17082 0

### 引用本文

Tian He, Li Daojing, and Qi Chunchao. Millimeter-wave human security imaging based on frequency-domain sparsity and rapid imaging sparse array architecture[J]. Journal of Radars, 2018, 7(3): 376-386. DOI: 10.12000/JR17082.

### 文章历史

(中国科学院电子学研究所微波成像技术重点实验室    北京    100190)
(中国科学院大学    北京    100190)
(华讯方舟科技有限公司    深圳    518000)

Millimeter-wave Human Security Imaging Based on Frequency-domain Sparsity and Rapid Imaging Sparse Array Architecture
Tian He①②, Li Daojing, Qi Chunchao
(Science and Technology on Microwave Imaging Laboratory, Institute of Electronics, Chinese Academy of Sciences, Beijing 100190, China)
(University of Chinese Academy of Sciences, Beijing 100190, China)
(China Communication Technology Co., Ltd., Shenzhen 518000, China)
Foundation Item: The National Natural Science Foundation of China (61271422)
Abstract: This paper examines the processing of millimeter-wave imaging data based on sparse sampling and sparse array design for the rapid imaging of human security data. First, based on the cylindrical scanning imaging model, the Barker code-based randomly sparse sampling method is employed to reduce the scanning time. Then, a three-dimensional imaging algorithm based on interferometry and compressed sensing in the frequency domain is proposed, with sparse representation of the image in the frequency domain after interferometry and Compressed Sensing measurement model, to recover the image frequency spectrum, thereby implementing human security image reconstruction via sparse sampling. Real data processing results indicated that the proposed method could obtain image resolution and performance similar to those of complete samples and that the image correlation coefficients before and after sparse sampling were better than 0.9, with 50% time/data reduction. Furthermore, based on the Barker codes and multistatic work mode, a sparse array architecture for rapid imaging was designed with a sparse rate of 94.6% and the guarantee of imaging quality. The proposed method was found to considerably increase the passage rate and reduce the amount of radiation unit and system complexity, marking its application significance and market prospect in security clearance.
Key words: Millimeter wave imaging    3D imaging    Sparse sampling    Sparse array    Interferometry    Human security
1 引言

2003年，美国L3通信公司研制了ProVision毫米波人体扫描仪[4]。该设备工作频率为 24～30 GHz，所采用的图像处理技术可获得5 mm横向分辨率，目前在美国、欧洲及亚洲等部分机场均有使用。2011年，德国埃尔兰根-纽伦堡大学微波光子研究所搭建了多基地毫米波快速成像系统[5,6]。该原型系统工作频率为70～80 GHz，所获得的成像结果横向分辨率为2 mm；采用一个2 m×1 m大小的2维平面稀疏阵列，平板式数据获取时间可达到50 ms以下，可实现人体目标实时成像。

2 成像模型及信号处理 2.1 柱面扫描3维成像模型

 图 1 毫米波人体安检柱面扫描3维成像模型 Fig.1 The cylindrical scanning 3-D imaging model of a millimeter wave human security system

 图 2 巴克码[1110010]的自相关函数图和方向图 Fig.2 The Autocorrelation and Beam pattern diagrams for Barker code of [1110010]
2.2 回波信号处理

$g({\theta _a},{k_r})$ ${F_\sigma }({k_r},{\theta _a},{k_z})$ 关于 ${\theta _a}$ 做1维傅里叶变换，则式(6)可写为：

3 频域稀疏3维成像算法

${{σ}} {\rm{ = }}{{{F}}^{{\rm{ - 1}}}}{{α}}$ , F–1为逆傅里叶基，则

4 柱面扫描实际数据处理 4.1 系统参数

 图 3 柱面扫描型成像系统和测试目标 Fig.3 The cylindrical scanning imaging system and testing objectives

 图 4 柱面回波数据稀疏方式示意图 Fig.4 The sparse sampling method of cylindrical echo
4.2 成像结果

 图 5 满采样条件下的人体安检直接成像结果 Fig.5 The direct imaging result under complete sampling

 图 6 等间隔稀疏采样条件下的人体安检直接成像结果 Fig.6 The direct imaging result under uniformly spaced sparse sampling
 图 7 基于巴克码稀疏采样条件下的人体安检直接成像结果 Fig.7 The direct imaging result under Barker code sparse sampling

 图 8 干涉处理前后图像3维频谱 Fig.8 The 3-D frequency spectrum before and after interferometry
 图 9 基于干涉处理和频域CS 3维成像算法的图像重建结果 Fig.9 The reconstruction of 3-D imaging algorithm based on interferometry and CS in frequency domain
4.3 误差分析

5 用于快速安检成像的平面稀疏阵列设计

5.1 2维稀疏阵列布局

 图 10 2维稀疏阵列中最小单元结构示意图 Fig.10 The structure of minimum unit in 2-D sparse array

 图 11 毫米波快速成像2维稀疏阵列布局示意图 Fig.11 The 2-D sparse array structure for millimeter wave fast imaging
5.2 指标分析

(1) 成像分辨率

(2) 稀疏度

6 结论