隧道工程人员身份安全识别  

Recognition algorithms of security identification system for tunnel engineering

作　　者：王君 徐铭江 梁薇薇 陈定安 孙晔然 WANG Jun;XU Mingjiang;LIANG Weiwei;CHEN Dingan;SUN Yeran(School of Electrical and Computer Engineering,Nanfang College of Sun Yat-sen University,Guangzhou 510970,China;School of Automation,Chongqing University of Posts and Telecommunications,Chongqing 400065,China;Guangzhou Hengtong Zhilian Technology Limited Company,Guangzhou 510630,China;School of Geography and Planning,Sun Yat-sen University,Guangzhou 510970,China;Department of Geography,College of Science,Swansea University,sgeti,Swansea SA28PP,UK)

机构地区：[1]中山大学南方学院电气与计算机工程学院,广东广州510970 [2]重庆邮电大学自动化学院,重庆400065 [3]广州恒通智联科技有限公司,广东广州510630 [4]中山大学地理科学与规划学院,广东广州510970 [5]斯旺西大学理学院地理系,斯旺西SA28PP

出　　处：《电子设计工程》2021年第14期145-148,153,共5页Electronic Design Engineering

摘　　要：铁路隧道工程环境复杂,实时获得人员身份信息,有利于险情下救援。文中针对在特殊隧道下光线不足、粉尘附着、遮挡时要求提高识别率的目的,通过融合多种生物特征的复合识别技术手段:分别提取面部、虹膜、声音特征,得出特定特征向量并由二维离散余弦变换,除掉会干扰到结果的信息;然后,通过决策树支持向量机(Decision Trees Support Vector Machine,DT-SVM)分类器匹配,实现工程人员身份识别。通过实验可知,该方法不仅降低了在计算时的难度,而且在信噪比为15~35 dB时,识别率在97%以上,提高了身份识别的准确性。The environment of the tunnel engineering is complex,monitoring the identity on line,it is convenient for rescue in danger.The recognition of traditional single recognition is unsuccessful,such as encounters light,dust,occlusion.For the above problems,this papaper proposes a recognition with multiple biological features.It results a new feature vector from the features of human face,iris and fingerprint.Then,matches the new features and completes the identification.Through the 2 D-DCT(Two Dimensional Discrete Cosine Transform)reduces lots of useless feature,and the recognition is accurately achieved by the DT-SVM classifier(Decision Trees Support Vector Machine).The results shows:this method reduces the complexity,the recognition rate is above 97% with the SNR(Signal to Noise Ratio)between15 dB and 35 dB,it improves the accuracy of recognition.

关 键 词：隧道工程 生物特征融合 二维离散余弦变换 决策树支持向量机分类器 

分 类 号：TN957.52[电子电信—信号与信息处理]