基于嵌入式虹膜身份识别系统研究

发布时间：2018-06-10 03:25

本文选题：虹膜识别 + Linux操作系统　；参考：《重庆理工大学》2017年硕士论文

【摘要】：生物特征识别技术是根据人体本身拥有的特征或者行为来识别身份的技术,虹膜识别属于生物特征识别技术中的一种。虹膜中含有非常丰富的纹理信息,并且具有终生稳定的特性,使得它具有很好的研究价值和发展潜力。在安全性越来越被重视的当今社会,虹膜识别技术也在逐渐地普及到人们的生活中。为了适应当今社会的身份认证需求,必须寻求一种功耗低、成本低、稳定性好、识别率高、可移动性强的生物识别系统,本文研究了一种基于嵌入式的虹膜识别系统。主要内容包括以下几个方面:1、研究了一款嵌入式的虹膜识别系统,采用的是基于ARM9内核的mini2440开发板作为虹膜识别系统的硬件模块,在开发板上移植了linux操作系统,完成USB、串口、LCD液晶显示等相关驱动。并利用Qt作为集成开发环境来设计虹膜识别的图形界面,使之具有良好的人机交互界面。2、在虹膜定位算法上,综合考虑Hough变换和微积分圆检测器的优缺点,分别利用这两种算法的改进方法来定位虹膜的边缘,经过大量的实验,表明了本文方法的定位效果很好,同时也具有一定的抗干扰能力。3、针对传统的Hamming距离特征匹配存在移位比对次数较多的问题,本文提出了一种改进的虹膜特征匹配算法,通过自动控制移位比对的次数,达到合理地减少比对次数的效果,缩短了虹膜特征匹配的时间,改善了虹膜识别系统的性能。4、利用中科院虹膜数据库CASIA1.0,在PC机和本文所研究的嵌入式虹膜识别系统平台上进行了大量实验,得到了虹膜识别所耗费的时间和识别的正确率。通过与传统方法的对比,本文方法在识别时间上较短,精度更高,证实了本文所设计的虹膜识别系统具有较好的识别性能。经过大量的实验测试,本文所研究的嵌入式虹膜识别系统具有识别率高、抗干扰能力强、成本低、可移动性好等优点,在实际的应用中具有一定的价值。
[Abstract]:Biometric recognition technology is a kind of identification technology based on the characteristics or behaviors of human body. Iris recognition is one of the biometric identification techniques. The iris contains rich texture information and has the characteristics of lifetime stability, which makes it has good research value and development potential. In today's society, where more and more attention has been paid to security, iris recognition technology is becoming more and more popular in people's lives. In order to meet the needs of identity authentication in today's society, we must seek a biometric recognition system with low power consumption, low cost, good stability, high recognition rate and high mobility. In this paper, a kind of iris recognition system based on embedded system is studied. The main contents include the following several aspects: 1, this paper studies an embedded iris recognition system. The mini2440 development board based on ARM9 kernel is used as the hardware module of iris recognition system, and the linux operating system is transplanted on the development board. Complete USB, serial LCD LCD and other related drivers. QT is used as the integrated development environment to design the graphical interface of iris recognition, so that it has a good human-computer interface. In the iris localization algorithm, the advantages and disadvantages of Hough transform and calculus circle detector are considered synthetically. The improved methods of these two algorithms are used to locate the edge of iris respectively. A lot of experiments show that the localization effect of this method is very good. At the same time, it also has some anti-interference ability. 3. In view of the problem of the traditional hamming distance feature matching, the paper proposes an improved iris feature matching algorithm, which can automatically control the frequency of shift matching. To achieve the effect of reasonably reducing the times of comparison and shortening the time of iris feature matching, The performance of iris recognition system is improved .4. by using CASIA 1.0, a large number of experiments are carried out on PC and the embedded iris recognition system platform studied in this paper. The time spent in iris recognition and the correct rate of iris recognition are obtained. Compared with the traditional method, the method in this paper has shorter recognition time and higher accuracy, which proves that the iris recognition system designed in this paper has better recognition performance. After a lot of experimental tests, the embedded iris recognition system studied in this paper has the advantages of high recognition rate, strong anti-interference ability, low cost, good mobility and so on. It has certain value in practical application.
【学位授予单位】：重庆理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41

【参考文献】