红外, 2011, 32 (12): 23, 网络出版: 2011-12-20
基于阈值分割和边缘检测的枪支THz图像识别
Feature Recognition of Pistol in Terahertz Image Based on Threshold Segmentation and Edge Detection
太赫兹成像 安全检查 图像处理 特征提取 图像识别 terahertz imaging security check image processing feature extraction image recognition
摘要
针对太赫兹(THz)图像的特点,采用阈值分割和边缘检测等数字图像处理技术将枪支分割出 来,然后提取枪支的有效特征,并对人体携带枪支的THz图像进行自动识别和判断。若发现有危险品或可疑物 品,则发出警报信号。实验结果表明,该方法能够有效地将人体隐蔽携带的枪支类危险品同钢笔、钥匙、挂饰等人 体通常携带的其它物品区别开来,实现了安全检查中对枪支**类违禁品THz图像的自动识别。
Abstract
According to the characteristics of THz images, some digital image processing methods such as threshold division and edge detection are used to segment the pistol from a THz image. Then, the effective signatures of the pistol are extracted and automatic recognition and judgment are implemented on the image. If hazardous or suspicious goods are found, an alarm signal is generated. The experimental result shows that this method can effectively discriminate the pistols concealed in clothes from other metal goods such as pens, keys and ornaments which people usually carry. It can be used to automatically identify the contraband goods such as pistols in the THz images in a security check.
冯艳平, 王徽. 基于阈值分割和边缘检测的枪支THz图像识别[J]. 红外, 2011, 32(12): 23. FENG Yan-ping, WANG Hui. Feature Recognition of Pistol in Terahertz Image Based on Threshold Segmentation and Edge Detection[J]. INFRARED, 2011, 32(12): 23.