基于手指静脉的身份识别方法研究
VIP免费
I
摘 要
传统的身份认证方法如钥匙,卡片,密码,PIN 码等。存在物品遗失或密码
容易被遗忘的缺点,一旦别人通过非法途径获得了标示物或者密码就会拥有相同
的权利。据 Master Card 公司估计,每年约有 4.5 亿美元的信用卡诈骗案件发生,
其中就包括利用遗失和被盗的信用卡犯罪,但是如果销售场所可以准确无误的识
别持卡人身份就可以大大减少这类案件的发生。因此,一种高安全性、高可靠性
的身份识别系统—基于手指静脉特征的身份识别系统—应运而生。
本文在 MATLAB 平台的基础上实现了一系列算法,从而达到利用人的手指静脉
特征进行身份识别的目的。 本文将以如下线索展开:
首先,设计并实现了手指静脉图像采集系统,实验证明,该系统能满足手指
静脉图像采集的要求;根据手指静脉图像的特点,设计并实现了图像的平滑去噪、
锐化和二值化算法,从而将手指静脉区域和背景区域较清晰的分离开来。在锐化
部分,改进了传统的高频强调滤波方法,而采用将高频强调滤波和直方图均衡化
相结合的方法,达到了增强静脉和背景对比度的较理想的效果。在二值化处理部
分,根据手指静脉不同区域灰度差别较大的特点,提出了一种分区域处理与形态
学图像处理相结合的方法,分离出了静脉和背景,并且为了减少系统响应时间,
引入了将图像归一化的思想;最后,将手指静脉纹路进行了细化,并用模板匹配
识别的方法实现了整个系统的设计初衷。
本文从分析手指静脉图像的特点出发,基于 MATLAB 平台实现了图像归一化、
去噪、增强、二值化和细化,并用模板匹配的方法达到识别的目的。实验结果表
明,用本文方法可以达到比较理想的身份识别效果。
关键词:手指静脉 身份识别 高频强调滤波 图像细化 模板
匹配
III
ABSTRACT
The traditional identification methods, such as keys, cards, password, user name
and etc, have some disadvantages, for example, the keys and cards are easily lost and
the password and user name will be forgot or somebody else may get to know it.
Once somebody else get these materials or passwords, he has the same right to do
things just like the owners. According to Master Card Company, there are about 450
million dollars fraud using lost or stolen credit cards every year, and if we can identify
the owners accurately, we can decrease the number of this kind of fraud. So we need a
more safety and more reliability system —Finger -vein identification system to
indentify different people.
This paper realized a series of algorithms using MATLAB language, and get to
the target of indentifying people with finger-vein. We will begin to introduce our
finger-vein identification system in the following steps.
First of all, we designed and assembled the finger-vein image collection system,
and it is proved by experiment that it can satisfy our demands. Secondly, according to
the characteristics of the finger-vein image, we adopted a series of methods to enhance
the contrast of the image in order to separate the finger-vein areas from the background
areas. The method consists of three steps: denoising, contrast enhancement and image
binarization. In denoising, considering the relationship between gray levels in the
adjacent areas of the finger-vein image, we adopted the Gradient Inverse weighted
smoothing method. In contrast enhancement, we adopted a method which combined
the traditional high frequency stress filtering algorithm together with the histogram
equalization. With this method, the contrast of the finger-vein area and the background
area has been enhanced significantly. During the binarization process, after taking the
differences of the gray levels between the different areas of the finger-vein image into
consideration, we proposed a method which is based on dividing the image into several
segments. Our results show that this set of means is quite competent to separate the
finger-vein areas from the background areas. What’s more, in order to reduce the
processing time of the system, we introduced the idea of reducing the size of the
finger-vein image to a smaller probable size. At last, we refined the finger-vein and
using the template matching method to realize the target of the whole system.
This paper analyzed the characteristics of finger-vein image, and based on
III
MATLAB platform, we realized image normalization, denoising, enhancement,
binarization and thinning, and used the template matching method to achieve the
purpose of identification. Experimental results show that this method can be used to
achieve the ideal effects of identification.
Key words: Finger-vein,Personal identification,High frequency
stress filtering,Image thinning,Template match
目 录
中文摘要
ABSTRACT
第一章 绪 论 .........................................................1
§1.1 引 言 ........................................................ 1
§1.2 几种典型的生物特征识别技术及其性能比较 ....................... 1
§1.2.1 基于生理特征的生物识别技术 ............................... 2
§1.2.1.1 人脸识别 ..............................................2
§1.2.1.2 指纹识别 ..............................................2
§1.2.1.3 手形识别 ..............................................3
§1.2.1.4 基因(DNA)识别 .......................................3
§1.2.1.5 虹膜识别 ..............................................4
§1.2.2 基于行为特征的生物识别技术 ............................... 4
§1.2.2.1 语音识别 ..............................................4
§1.2.2.2 签名识别 ..............................................5
§1.2.2.3 步态识别 ..............................................5
§1.2.3 几种典型的生物识别技术性能比较 ........................... 5
§1.3 手指静脉识别技术简介 ......................................... 6
§1.4 手指静脉识别技术国内外研究现状 ............................... 7
§1.5 手指静脉识别技术的研究内容 ................................... 8
§1.6 生物特征识别技术市场前景 ..................................... 8
§1.7 本章小结 ..................................................... 9
第二章 手指静脉图像的采集 ........................................... 11
§2.1 数字图像处理和 MATLAB 工具箱的背景知识简介 .................. 11
§2.2 手指静脉识别原理 ............................................ 12
§2.3 基于 MATLAB 的手指静脉识别系统的系统组成 ..................... 12
§2.4 系统主要硬件组成 ............................................ 13
§2.4.1 红外光源的选择 ........................................... 13
§2.4.2 成像设备的选择 .......................................... 15
§2.5 采集装置的实现和图像的采集 .................................. 15
§2.6 本章小结 .................................................... 16
第三章 手指静脉图像的预处理 ......................................... 17
§3.1 手指静脉图像的边缘提取与定位 ................................ 17
§3.1.1 几种典型的边缘提取算法比较 .............................. 17
§3.1.1.1 ROBERTS 算子 ........................................18
§3.1.1.2 SOBEL 算子 ..........................................18
§3.1.1.3 LAPLACIAN 算子 ......................................19
§3.1.1.4 几种算子性能优劣的比较 ...............................20
§3.1.2 手指静脉图像的截取 ...................................... 20
§3.2 手指静脉图像尺寸归一化 ...................................... 21
§3.2.1 归一化常见方法介绍 ...................................... 21
§3.2.1.1 最近邻插值法 ........................................22
§3.2.1.2 双线性插值法 ........................................22
§3.2.2 尺寸归一化 .............................................. 23
§3.3 本章小结 .................................................... 24
第四章 手指静脉图像的后处理 ......................................... 25
§4.1 采集到的图像产生模糊的原因 .................................. 25
§4.2 手指静脉图像的平滑去噪 ...................................... 25
§4.2.1 简单平均化滤波去噪 ...................................... 25
§4.2.1.1 局部平均滤波 .........................................26
§4.2.1.2 局部加权平均滤波 .....................................26
§4.2.2 不损害画质的平滑化 ...................................... 26
§4.2.2.1 中值滤波 .............................................27
§4.2.2.2 基于梯度倒数权重法的平滑去噪 .........................27
§4.3 将高频强调滤波和直方图均衡化相结合的图像锐化 ................ 28
§4.3.1 高通滤波 ................................................ 28
§4.3.1.1 空域高通滤波器 .......................................28
§4.3.1.2 频域高通滤波 .........................................29
§4.3.2 高频强调滤波 ............................................ 30
§4.4 基于分区域二值化与形态学图像处理的图像分离 .................. 31
§4.5 本章小结 .................................................... 32
第五章 手指静脉图像的匹配和识别 ..................................... 33
摘要:
展开>>
收起<<
I摘要传统的身份认证方法如钥匙,卡片,密码,PIN码等。存在物品遗失或密码容易被遗忘的缺点,一旦别人通过非法途径获得了标示物或者密码就会拥有相同的权利。据MasterCard公司估计,每年约有4.5亿美元的信用卡诈骗案件发生,其中就包括利用遗失和被盗的信用卡犯罪,但是如果销售场所可以准确无误的识别持卡人身份就可以大大减少这类案件的发生。因此,一种高安全性、高可靠性的身份识别系统—基于手指静脉特征的身份识别系统—应运而生。本文在MATLAB平台的基础上实现了一系列算法,从而达到利用人的手指静脉特征进行身份识别的目的。本文将以如下线索展开:首先,设计并实现了手指静脉图像采集系统,实验证明,该系统能...
相关推荐
-
VIP免费2025-01-09 4
-
VIP免费2025-01-09 4
-
VIP免费2025-01-09 4
-
VIP免费2025-01-09 4
-
VIP免费2025-01-09 4
-
VIP免费2025-01-09 5
-
VIP免费2025-01-09 4
-
VIP免费2025-01-09 4
-
VIP免费2025-01-09 4
-
VIP免费2025-01-09 4
作者:陈辉
分类:高等教育资料
价格:15积分
属性:50 页
大小:920.16KB
格式:PDF
时间:2024-11-19