计算机工程与应用 ›› 2012, Vol. 48 ›› Issue (28): 193-196.

• 图形、图像、模式识别 • 上一篇    下一篇

一种安全检查的无源毫米波图像特征提取方法

秦文杰,张光锋,娄国伟   

  1. 南京理工大学 电子工程与光电技术学院,南京 210094
  • 出版日期:2012-10-01 发布日期:2012-09-29

Feature extraction method of PMMW radiation image based on security inspection

QIN Wenjie, ZHANG Guangfeng, LOU Guowei   

  1. School of Electronic Engineering and Optoelectronic Technology, Nanjing University of Science and Technology, Nanjing 210094, China
  • Online:2012-10-01 Published:2012-09-29

摘要: 毫米波穿透力强且具有全天时、全天候工作能力,在反恐安检方面具有广阔的应用前景。针对藏匿物品的安全检查获取的毫米波辐射图像,与小波软硬阈值预处理滤除噪声算法相比,提出一种复合结构形态学自适应滤波算法对毫米波辐射图像进行了滤波处理,然后对处理后图像进行边缘检测与特征分析。实验结果表明,复合结构形态学自适应滤波算法能有效地去除毫米波辐射图像中的噪声,提取的图像边缘特征符合藏匿物品的几何特征。

关键词: 毫米波辐射计, 安全检查, 小波去噪, 形态学滤波, 边缘检测

Abstract: Because of its strong penetration and work ability of all-time and all-weather, MMW radiation has vast application prospects in fields of counter-terrorism security inspection. Compared with the usual soft and hard threshold wavelet de-noising algorithms, a complex morphology self-adaptive filtering algorithm is proposed according to the MMW radiation image gotten from security inspection for canceled objects. Based on the algorithms above, the MMW radiation image is pre-processed. The edge of the image is detected and the feature of the image is extracted. The results show that the complex morphology self-adaptive filtering algorithm can remove more image noise than soft threshold wavelet de-noising algorithm and the features of edges extracted are consistent to the real geometric characteristics of canceled objects.

Key words: MMW radiometer, security inspection, wavelet de-noising, morphology filtering, edge detection