【作者】 罗涛；

【导师】 徐端全； 傅敏；

【作者基本信息】 华中科技大学 ， 电子与通信工程， 2017， 硕士

【摘要】 身份证是证明持有人身份的有效证件,在机场、火车站、酒店、手机营业厅、银行等场所,均需要用到第二代居民身份证进行身份认证。在认证的过程中,通过识读身份证芯片完成身份证的真伪鉴别和个人信息的自动录入,并通过人眼判断身份证与持卡人的一致性。当前身份证的认证方式效率不高,鉴伪有漏洞,存在人证不一致的问题。针对这些问题,本文基于相关硬件平台,设计并开发了一套融合了身份证图像采集、芯片信息读取、身份证鉴伪和持卡人识别的软件系统。首先,设计了身份证图像采集上位机程序,接收CIS传感器采集的身份证图像数据,针对采集到的身份证原始图像颜色偏暗、存在条纹噪声等情况,本文设计了基于CIS传感器的身份证图像亮度补偿校正预处理方法,并通过边缘检测和图像缩放的方法对身份证图像标准化;其次,本文使用了公安部认证的非接触式二代证阅读器二次开发模块对身份证读卡程序进行开发设计,读取身份证内置芯片信息。本文还实现了基于特征提取和SVM分类器的身份证字符识别方法,并根据身份证字符的特点,设计了不同的分类器对身份证上不同的字符信息进行分类识别,大大提升了身份证字符识别的速度,通过比对身份证芯片信息和识别的文本信息,鉴别身份证真伪;最后,本文利用从身份证图像中定位分割出的人脸头像,通过摄像头实时检测人脸信息,识别身份证持卡人是否身份证本人。本文使用C++语言实现了上述算法,并结合MFC完成了相关软件的开发。经过实验测试,本系统采集的身份证图像清晰度较高,芯片信息读取稳定,对身份证图像文本信息有较高的识别率,并且人脸检测效果好,系统具备人脸识别的功能。更多还原

【Abstract】 Identity-card is the valid document that prove the identity of the holder,in the airport,railway station,hotel,mobile phone business hall,bank and other place,are required to use the second generation of identity card for identity authentication.In the process of certification,we need through reading the identity card chip information to complete the identification of identity-card and automatic input personal information.After that,it’s necessary to identify the consistency of identity-card and cardholder by human eyes.The current identity card authentication method is not efficient,and loopholes exist when identify the authenticity of identity-card.Besides,there are problems of inconsistencies between Identity-card and it’s holder.In the view of above situations,this paper based on the relevant hardware platform,design and develop a software system that integrates ID card image acquisition,chip information reading,identity card authentication and cardholder identification.First of all,we designed ID card image acquisition PC program,receiving ID card image data collected by CIS sensor.Aiming at the dark color of the original image of the collected ID card and the presence of fringe noise,this paper designed a pre-processing method of brightness compensation image correction based on CIS sensor.And through the edge detection and image scaling method to standardize ID card image.Second,this article uses the Ministry of Public Security certified non-contact second-generation card reader secondary development module on the identity card reader to develop the design,read ID card built-in chip information.This paper also realizes the character recognition method based on feature extraction and SVM classifier,and according to the characteristics of ID characters,the design of different classifiers on the identity card on the different character information classification and identification,greatly enhance the ID card character recognition speed.By comparing the identity card chip information and identify the text information,identify the authenticity of identity-card.Finally,this paper location and split out the human face from ID card image,through the camera real-time detected human face head,identify whether the identity-card holder matching the identity card.In this paper,the C ++ language is used to implement the above algorithm,and the related software is developed with MFC.After the experimental test,the system has a high degree of image resolution,chip information read stability,the identity-card image text information has a higher recognition rate,and face detection effect,with face recognition function.更多还原