系统工程与电子技术

• 软件、算法与仿真 • 上一篇    下一篇

主动毫米波图像的人体携带危险物检测研究

杜琨1, 王威2, 年丰2,3, 陈炜1, 胡锋杰1   

  1. 1. 北京航空航天大学电子信息工程学院, 北京 100191; 2. 北京无线电计量测试研究所,
    北京 100854; 3. 计量与校准技术重点实验室, 北京 100854
  • 出版日期:2016-05-25 发布日期:2010-01-03

Concealed objects detection in active millimeter-wave images

DU Kun1, WANG Wei2, NIAN Feng2,3, CHEN Wei1, HU Feng-jie1   

  1. 1. School of Electronics and Information Engineering, Beihang University, Beijing 100191, China;
    2. Beijing Institute of Radio Metrology and Measurement, Beijing 100854, China;
    3. Science and Technology on Metrology and Calibration Laboratory, Beijing 100854, China
  • Online:2016-05-25 Published:2010-01-03

摘要:

近距离主动毫米波人体成像是检测藏匿在人体衣物下危险品的一种有效手段。针对主动毫米波图像分辨率高,人体不同部位呈现不同特性,以及人体、其他携带物和危险物特性相互影响和干扰、危险物检测困难的特点,提出了一种主动毫米波图像多危险物检测方法;通过基于人体特征的区域划分,采用图像区域化的自适应阈值图像分割、结合数学形态学的二次提取方法,实现了主动毫米波图像人体携带多危险物的检测和提取。对实际获取的图像进行了多危险物检测实验,检测率为84.13%,虚警率为6.61%。仿真实验结果表明了所提方法的有效性和实用性。

Abstract:

Active millimeter-wave near-filed human imaging is a means for concealed objects detection. A method of concealed objects detection in the usage of active millimeter-wave images is presented as a result of image characteristics which include high resolution, characteristics varying in different parts of the human, imaging influenced among human, concealed objects and other objects, and difficulties in concealed objects detection. By dividing images according to human regional characteristics, this method applies image adaptive threshold segmentation and re-extraction with the mathematical morphology to detect concealed objects. The experiments of concealed objects on images of actual acquisition are conducted with a result of accurate rate 84.13% and false alarm rate 6.61%, illustrating the effectiveness and practicality of the proposed method.