基于阈值分割的THz图像刀具识别

Feature Recognition of Knife from Terahertz Image Based on Threshold Segmentation

本文将数字图像处理技术与智能化监控报警系统相结合,针对人体隐蔽携带刀具的太赫兹辐射成像的图像,采用阈值分割的方法将刀具分割出来,然后提取刀具的有效特征,将其识别出来,从而判断人体是否隐蔽携带了刀具类危险品。如发现为危险品或可疑物品就发出警报信号,以便监控人员进行进一步的安全检查。实验证明,该方法可以将刀具同钢笔、打火机等一些金属物品区分出来。

This article combines the digital image processing with the intelligent monitor alarm system. We use threshold division to extract knives features from terahertz image of body with hidden knives in order to judge if the person takes knives. Alarm signal will be sent when dangerous or suspicious goods are found. Then monitor staff can do the future safety examination. The experimental results prove that the proposed method can identify the knives from other metal items such as pens or lighters.