纸张残留指纹信息成像系统的研究
摘要
随着信息技术的高速发展，信息安全已越来越受到人们的普遍和高度重视。特别是安全部门对文件的保密工作尤为重视，对机密文件的审阅人有严格的限制。本文详细叙述了纸张上残留的痕迹（主要是汗潜指纹）的检测原理和工作过程。提取汗潜指纹通常采用的方法是用灰粉或墨粉进行处理，但在渗透性材料上，如纸张、硬纸板等最合适的方法是采用碘熏法。这些方法都会对检样造成不同程度上的损坏，为了保持原检样的完整性，本文提出一种对纸张上的汗潜指纹进行检测与成像的新方法。通过检测纸张有无痕迹处的阻抗，并将此阻抗信号处理后，输入到计算机后将其转换成图像信息，再利用Visual C++算法实现图像增强。主要采用了阈值变换、灰度拉伸、窗口变换、小波变换、图像的平滑以及锐化等操作，使处理后的图像质量符合进一步分割和识别的要求。
在本文的第一章中，我们介绍了整个系统的方案论证和整体思路，并给出了可行性分析和项目研究的意义。
在第二章中对检测的原理及相关电路进行了详细说明。由于纸张上有无残留痕迹处的阻抗不同，因此可以通过测试纸张有无汗潜指纹纹理处的阻抗，来实现检测纸张上残留痕迹的目的。通过检测取样电阻Ro12两端的电压值，将此电压信号经线性放大器放大、A/D转换等处理后，传送到微机中，利用微机对所采集的数据进行处理，将处理的结果在监视器上显示，并以图像文件（JPG 或BMP格式）方式存储，从而达到检测痕迹的目的。显然，对数据的采集部分要求一定的灵敏度，为此我们考虑检测时的直流参数，有：
(01)
其中：R1+R、R2+R是有无痕迹处纸张阻抗，R1、R2是有无痕迹处纸张变化阻抗，RO1、RO2是（同一个）取样电阻。
事实上有：R>>R1、R2、RO1、RO2
(1). 当取样电阻采用线性电阻时，即RO1=RO2=RO，则有： (02)
(2). 当取样电阻采用非线性电阻时：
(03)
因此，采用非线性电阻作为取样电阻，有利于测试。这对于提高显示灵敏度至关重要。研制非线性取样器件（电路）及与之配套的前置放大器是决定该设备显示灵敏度的关键。实际电路设计时采用了恒流源电路，激励源采用了正弦交流信号源和直流信号源两种激励方式。
采集的原始数据与位图格式具有一一对应性，由于系统及纸张湿度的不确定性、纸张厚度的不均匀性以及外界的干扰等因素，使得原始数据中背景噪声很大，有必要对背景信号和随机噪声进行最大限度的抑制，以保证高质量的再现目标信息。因此需要对数据进行必要的处理。把区域内的汗潜指纹信息看成目标图像,而区域的纸张看作图像的背景,如果由于纸张湿度、厚度的不均匀可能导致背景信号的变化过大，当这种变化接近目标信号时，对目标信号的识别和分辨是非常不利的，我们没有合适的方法改变纸张的厚度，因此只能用数据处理的方法最大限度的抑制背景噪声。对图像信号的取样存在噪声，这种噪声为白噪声，随机出现。噪声过大就会造成背景图像的失真，必须加以滤除，由于是随机噪声的影响，因此可以采取多幅图像的平均值作为背景。这部分的数据处理可在硬件系统中处理，将处理的背景信息存储在背景存储器中，其实际采集数据与背景数据相减。由于纸张的材质、湿度的不同，因此要求对不同的纸张测试时都要求进行背景信号采集。实际电路中采取了差分测量电路，即对每点的测量值均与某对应点的平均值相减。在本文的第三章中，不但给出了这种处理的理论基础，而且对盲解卷积图像恢复的理论进行了简要介绍。
第四章中，重点介绍了图像增强的方法和这种方法对本系统的贡献。图像增强是指按特定的需要突出一幅图像中的某些信息，同时，削弱或去除某
些不需要的信息的处理方法。其主要目的是使处理后的图像对某种特定的应用来说，比原始图像更适用。因此，这类处理是为了某种应用目的而去改善图像质量的。处理的结果使图像更适合于人的视觉特性或机器的识别系统。应该明确的是增强处理并不能增强原始图像的信息，其结果只能增强对某种信息的辨别能力，而这种处理有可能损失一些其他信息。
图像增强技术主要包括直方图修改处理，图像平滑化处理，图像尖锐化处理及彩色处理技术等。在实用中可以采用单一方法处理，也可以采用几种方法联合处理，以便达到预期的增强效果。
图像增强技术基本上可分成两大类：—类是频域处理法，—类是空域处理法。频域处理法的基础是卷积定理。它采用修改图像傅里叶变换的方法实现对图像的增强处理。如果原始图像是，处理后的图像是，而是处理系统的冲激响应，那么，处理过程可由下式表示：=* (04)
其中*代表卷积。如果，，，分别是，，的傅里叶变换。那么，上面的卷积关系可表示为变换域的乘积关系，即= (05)
式中，为传递函数，在增强问题中，是给定的原始数据，经傅里叶变换可得到。选择合适的，使得由式=得到的比在某些特性方面更加鲜明、突出，因而更加易于识别、解译。
空域法是直接对图像中的像素进行处理，基本上是以灰度映射变换为基础的。所用的映射变换取决于增强的目的
With the rapid development of information technology, information security has been given increasing attention. The security departments, in particular, have attached greater importance to the document secrecy and they impose strict restrictions on the readers of the secret documents. This article elaborates on the testing principles and working procedures of the remaining trace on paper (mainly invisible fingerprints of sweat). Now the generally accepted method to handle and collect invisible fingerprints of sweat is by means of lime or ink powder. And the most proper way on such penetrating material carriers as paper and cardboard is steaming iodine approach. However all the above-mentioned methods will damage the testing samples on varying degrees. In order to keep the original testing samples intact, a new method is proposed to test and get image of the invisible of fingerprints of sweat. The new approach is to test the difference on resistance between the remaining trace in paper and non-remaining trace, and input the tested data into computer. This approach makes use of Visual C++6.0 to convert the tested data into image and the Visual C++ computing method to obtain a keener image. This approach also adopts the operations as threshold transform, gray-scale stretch，window transform, image smoothing, image sharpening, so that the processed image can meet the requirements of further division and identity.
Chapter 1 mainly tells us about the whole plan demonstration and entire thinking of system. It also gives feasible analyze and item-study meaning.
It elaborates on principle of data collection and correlation circuits in Chapter 2. Because the resistances between the remaining trace and non-remaining trace are different in paper, we can test remaining trace on paper by testing the resistance of invisible fingerprints of sweat. By voltage of sampling resistance RO12 is tested. Next the voltage signal which is
magnified via linear amplifier. After converted to digital signal in A/D converter, the data is transmitted to PC. The data are deal with on PC. And the results are displayed on monitor and saved as image files (JPG or BMP format). We attain the examine trace aims. Evidently, the part of collecting data is required some sensitivities. So we consider the testing DC parameter.
(01)
Note: R1+R and R2+R are the paper resistances,
R1 and R2 are the paper-changing resistances,
Ro1 and Ro2 are the same sampling resistance.
In fact: R>>R1, R2, Ro1, Ro2
(1) When the sampling resistance uses the linear resistance Vi2,Ro1=Ro2=Ro, then:
(02)
(2) When the sampling resistance is non-linear resistance
(03)
Hence using non-linear resistance as sampling resistance is of advantage to testing. That is very important to improve display sensitivity. Researching non-linear sampling devices (circuit) and match preamplifier is the key to decide the equipment’s display sensitivity .In fact the circuit design adopts constant current source circuit, driving source adopts ways of sin AC signal and DC signal.
Sampling original data correspond with BMP format. But it is disturbed by the factors which are the humility uniform of system and paper, the uniformity of paper thickness and other outer interferences. This causes
background noise very noisy. So it is essential to restrain the background signal and random noise mostly which can make it sure to renew aim information in high quality. Data need disposing necessarily. The information about invisible fingerprints of surest in Si district is seen as aim image. While the paper in Si district is seen as the image background .The paper’s infirm of humidity and thickness maybe lead background signal to change too much. When this change approaches aim signal, it is disadvantage to discern aim signal. Now there is no suitable way to change paper thickness, so to restrain background noise only can uses the way of disposing data in most degree. The sampling of image signal lies in noise whic
在本文的第一章中，我们介绍了整个系统的方案论证和整体思路，并给出了可行性分析和项目研究的意义。
在第二章中对检测的原理及相关电路进行了详细说明。由于纸张上有无残留痕迹处的阻抗不同，因此可以通过测试纸张有无汗潜指纹纹理处的阻抗，来实现检测纸张上残留痕迹的目的。通过检测取样电阻Ro12两端的电压值，将此电压信号经线性放大器放大、A/D转换等处理后，传送到微机中，利用微机对所采集的数据进行处理，将处理的结果在监视器上显示，并以图像文件（JPG 或BMP格式）方式存储，从而达到检测痕迹的目的。显然，对数据的采集部分要求一定的灵敏度，为此我们考虑检测时的直流参数，有：
(01)
其中：R1+R、R2+R是有无痕迹处纸张阻抗，R1、R2是有无痕迹处纸张变化阻抗，RO1、RO2是（同一个）取样电阻。
事实上有：R>>R1、R2、RO1、RO2
(1). 当取样电阻采用线性电阻时，即RO1=RO2=RO，则有： (02)
(2). 当取样电阻采用非线性电阻时：
(03)
因此，采用非线性电阻作为取样电阻，有利于测试。这对于提高显示灵敏度至关重要。研制非线性取样器件（电路）及与之配套的前置放大器是决定该设备显示灵敏度的关键。实际电路设计时采用了恒流源电路，激励源采用了正弦交流信号源和直流信号源两种激励方式。
采集的原始数据与位图格式具有一一对应性，由于系统及纸张湿度的不确定性、纸张厚度的不均匀性以及外界的干扰等因素，使得原始数据中背景噪声很大，有必要对背景信号和随机噪声进行最大限度的抑制，以保证高质量的再现目标信息。因此需要对数据进行必要的处理。把区域内的汗潜指纹信息看成目标图像,而区域的纸张看作图像的背景,如果由于纸张湿度、厚度的不均匀可能导致背景信号的变化过大，当这种变化接近目标信号时，对目标信号的识别和分辨是非常不利的，我们没有合适的方法改变纸张的厚度，因此只能用数据处理的方法最大限度的抑制背景噪声。对图像信号的取样存在噪声，这种噪声为白噪声，随机出现。噪声过大就会造成背景图像的失真，必须加以滤除，由于是随机噪声的影响，因此可以采取多幅图像的平均值作为背景。这部分的数据处理可在硬件系统中处理，将处理的背景信息存储在背景存储器中，其实际采集数据与背景数据相减。由于纸张的材质、湿度的不同，因此要求对不同的纸张测试时都要求进行背景信号采集。实际电路中采取了差分测量电路，即对每点的测量值均与某对应点的平均值相减。在本文的第三章中，不但给出了这种处理的理论基础，而且对盲解卷积图像恢复的理论进行了简要介绍。
第四章中，重点介绍了图像增强的方法和这种方法对本系统的贡献。图像增强是指按特定的需要突出一幅图像中的某些信息，同时，削弱或去除某
些不需要的信息的处理方法。其主要目的是使处理后的图像对某种特定的应用来说，比原始图像更适用。因此，这类处理是为了某种应用目的而去改善图像质量的。处理的结果使图像更适合于人的视觉特性或机器的识别系统。应该明确的是增强处理并不能增强原始图像的信息，其结果只能增强对某种信息的辨别能力，而这种处理有可能损失一些其他信息。
图像增强技术主要包括直方图修改处理，图像平滑化处理，图像尖锐化处理及彩色处理技术等。在实用中可以采用单一方法处理，也可以采用几种方法联合处理，以便达到预期的增强效果。
图像增强技术基本上可分成两大类：—类是频域处理法，—类是空域处理法。频域处理法的基础是卷积定理。它采用修改图像傅里叶变换的方法实现对图像的增强处理。如果原始图像是，处理后的图像是，而是处理系统的冲激响应，那么，处理过程可由下式表示：=* (04)
其中*代表卷积。如果，，，分别是，，的傅里叶变换。那么，上面的卷积关系可表示为变换域的乘积关系，即= (05)
式中，为传递函数，在增强问题中，是给定的原始数据，经傅里叶变换可得到。选择合适的，使得由式=得到的比在某些特性方面更加鲜明、突出，因而更加易于识别、解译。
空域法是直接对图像中的像素进行处理，基本上是以灰度映射变换为基础的。所用的映射变换取决于增强的目的
With the rapid development of information technology, information security has been given increasing attention. The security departments, in particular, have attached greater importance to the document secrecy and they impose strict restrictions on the readers of the secret documents. This article elaborates on the testing principles and working procedures of the remaining trace on paper (mainly invisible fingerprints of sweat). Now the generally accepted method to handle and collect invisible fingerprints of sweat is by means of lime or ink powder. And the most proper way on such penetrating material carriers as paper and cardboard is steaming iodine approach. However all the above-mentioned methods will damage the testing samples on varying degrees. In order to keep the original testing samples intact, a new method is proposed to test and get image of the invisible of fingerprints of sweat. The new approach is to test the difference on resistance between the remaining trace in paper and non-remaining trace, and input the tested data into computer. This approach makes use of Visual C++6.0 to convert the tested data into image and the Visual C++ computing method to obtain a keener image. This approach also adopts the operations as threshold transform, gray-scale stretch，window transform, image smoothing, image sharpening, so that the processed image can meet the requirements of further division and identity.
Chapter 1 mainly tells us about the whole plan demonstration and entire thinking of system. It also gives feasible analyze and item-study meaning.
It elaborates on principle of data collection and correlation circuits in Chapter 2. Because the resistances between the remaining trace and non-remaining trace are different in paper, we can test remaining trace on paper by testing the resistance of invisible fingerprints of sweat. By voltage of sampling resistance RO12 is tested. Next the voltage signal which is
magnified via linear amplifier. After converted to digital signal in A/D converter, the data is transmitted to PC. The data are deal with on PC. And the results are displayed on monitor and saved as image files (JPG or BMP format). We attain the examine trace aims. Evidently, the part of collecting data is required some sensitivities. So we consider the testing DC parameter.
(01)
Note: R1+R and R2+R are the paper resistances,
R1 and R2 are the paper-changing resistances,
Ro1 and Ro2 are the same sampling resistance.
In fact: R>>R1, R2, Ro1, Ro2
(1) When the sampling resistance uses the linear resistance Vi2,Ro1=Ro2=Ro, then:
(02)
(2) When the sampling resistance is non-linear resistance
(03)
Hence using non-linear resistance as sampling resistance is of advantage to testing. That is very important to improve display sensitivity. Researching non-linear sampling devices (circuit) and match preamplifier is the key to decide the equipment’s display sensitivity .In fact the circuit design adopts constant current source circuit, driving source adopts ways of sin AC signal and DC signal.
Sampling original data correspond with BMP format. But it is disturbed by the factors which are the humility uniform of system and paper, the uniformity of paper thickness and other outer interferences. This causes
background noise very noisy. So it is essential to restrain the background signal and random noise mostly which can make it sure to renew aim information in high quality. Data need disposing necessarily. The information about invisible fingerprints of surest in Si district is seen as aim image. While the paper in Si district is seen as the image background .The paper’s infirm of humidity and thickness maybe lead background signal to change too much. When this change approaches aim signal, it is disadvantage to discern aim signal. Now there is no suitable way to change paper thickness, so to restrain background noise only can uses the way of disposing data in most degree. The sampling of image signal lies in noise whic
引文
李华. MCS51系列单片机实用接口技术.[M] 北京：北京航空航天大学出版社，1993. 499-526
Milan Snoka, Vaclav Hlavac, Roger Boyle, Image Processing, Analysis, and Machine Vision [M] Brooks/Cole, a division of Thomson Asia Pte Led, USA 57-118
何斌 马天予 王运坚 朱红莲. Visual C++ 数字图像处理.[M] 北京：人民邮电出版社，2001. 262-304
黄贤武 王加俊 李家华. 数字图像处理与压缩编码技术.[M] 成都：电子科技大学出版社，2000. 67-93
袁丁 傅一平 等. Visual C++精彩实例讲解. [M] 北京：机械工业出版社 2003 222-234
刘三满 指纹图像自动无损提取方法的研究 [J] 电脑开发与应用2001.1 13-15
牛仁朝 指纹识别系统的设计 [J] 电脑与信息技术1998.5 44-46
贾亮.纸张残留痕迹的检验与成像系统的研究 [硕士论文] 辽宁 东北大学.
刘小华 王燕生 指纹识别技术的发展 [J] 光学技术,1998,27(4):78-80.
周天泽 手纹趣谈.[J]知识就是力量, 2000,(1):45-46.
Pun T. Entropic thresholding，a new approach，CGIP, 1981，210~239
Brink AF thresholding of digital images using of two deminationel Entropies，Patlern Recognition.1992.25，803~808
Pratt W. pigital Image Processing.New York:Wiley. 1978.
Weszka J S .A survey of threshold selection techniques. Vision Graphics Image Process，1978，239~265
6 .Choi Y and Krishnapuram R. A robust approach to image enhancement based on fuzzy logic. IEEE Trans on Image Processing, 1997,6(6).
B. M. Mehtre, Fingerprint image analysis for automatic identification, Machine Vision and Application, 1993
CHENMao-Lin QIFei-Hu ADAPTIVE FILTER IN LOW QUALITY FINGERPRINT IMAGES [J] Infrared M illim. Waves December,2000
Max333datasheet.pdf http://www.maxim.com
Max132datasheet.pdf http://www.maxim.com
OP27datasheet.pdf http://www.icbase.com
INA116datasheet.pdf http://www.icbase.com
指纹识别系统及其应用http://www.china-fp.com /app_summarize/fingeryy.htm
指纹识别及相关技术的应用 http://www.china-fp.com /app_summarize/fingeryy_2.htm
周艳霞，计算机图像处理在射线探上中的应用 [D] 河北，河北工业大学 2001
Li T L，Sundarehan M K. Adaptive image contrast enhancement based on human visual properties[J] . IEEE Trans Medical Imaging,1994,13(4) :573-586.
Kim J Y, Kim L S. An advanced contrast enhancement using partially overlapped sub-block histogram equalization[J].IEEE Tram Circuits and
Video Technologp,2001 ,11(4) :475-484.
Chang D C，Wu W R. I mage contrast enhancement based on a histogram transformation of local standarddeviation[J].IEEE Trans Medical 1 maoino_1998_17f41-518-531.
于大勇等，频域迭代盲解卷积图像恢复方法及其算法实现[J].中国激光，2002.29(12):1101 1103.
吴颖谦等.基于小波变换的低对比度图像增强[J].红外与激光工程,2003.32(2):4 6.
李涛，小波理论及其在图像处理中的应用：[硕士学位论文].西安：西安电子科技大学.
许雷,郑筱祥,陈兴灿等.一种基于小波相位滤波及视觉非线性的医学图像自适应增强新方法[Ｊ].电子学报,1999,27(9):121 123.
Milan Snoka, Vaclav Hlavac, Roger Boyle, Image Processing, Analysis, and Machine Vision [M] Brooks/Cole, a division of Thomson Asia Pte Led, USA 57-118
何斌 马天予 王运坚 朱红莲. Visual C++ 数字图像处理.[M] 北京：人民邮电出版社，2001. 262-304
黄贤武 王加俊 李家华. 数字图像处理与压缩编码技术.[M] 成都：电子科技大学出版社，2000. 67-93
袁丁 傅一平 等. Visual C++精彩实例讲解. [M] 北京：机械工业出版社 2003 222-234
刘三满 指纹图像自动无损提取方法的研究 [J] 电脑开发与应用2001.1 13-15
牛仁朝 指纹识别系统的设计 [J] 电脑与信息技术1998.5 44-46
贾亮.纸张残留痕迹的检验与成像系统的研究 [硕士论文] 辽宁 东北大学.
刘小华 王燕生 指纹识别技术的发展 [J] 光学技术,1998,27(4):78-80.
周天泽 手纹趣谈.[J]知识就是力量, 2000,(1):45-46.
Pun T. Entropic thresholding，a new approach，CGIP, 1981，210~239
Brink AF thresholding of digital images using of two deminationel Entropies，Patlern Recognition.1992.25，803~808
Pratt W. pigital Image Processing.New York:Wiley. 1978.
Weszka J S .A survey of threshold selection techniques. Vision Graphics Image Process，1978，239~265
6 .Choi Y and Krishnapuram R. A robust approach to image enhancement based on fuzzy logic. IEEE Trans on Image Processing, 1997,6(6).
B. M. Mehtre, Fingerprint image analysis for automatic identification, Machine Vision and Application, 1993
CHENMao-Lin QIFei-Hu ADAPTIVE FILTER IN LOW QUALITY FINGERPRINT IMAGES [J] Infrared M illim. Waves December,2000
Max333datasheet.pdf http://www.maxim.com
Max132datasheet.pdf http://www.maxim.com
OP27datasheet.pdf http://www.icbase.com
INA116datasheet.pdf http://www.icbase.com
指纹识别系统及其应用http://www.china-fp.com /app_summarize/fingeryy.htm
指纹识别及相关技术的应用 http://www.china-fp.com /app_summarize/fingeryy_2.htm
周艳霞，计算机图像处理在射线探上中的应用 [D] 河北，河北工业大学 2001
Li T L，Sundarehan M K. Adaptive image contrast enhancement based on human visual properties[J] . IEEE Trans Medical Imaging,1994,13(4) :573-586.
Kim J Y, Kim L S. An advanced contrast enhancement using partially overlapped sub-block histogram equalization[J].IEEE Tram Circuits and
Video Technologp,2001 ,11(4) :475-484.
Chang D C，Wu W R. I mage contrast enhancement based on a histogram transformation of local standarddeviation[J].IEEE Trans Medical 1 maoino_1998_17f41-518-531.
于大勇等，频域迭代盲解卷积图像恢复方法及其算法实现[J].中国激光，2002.29(12):1101 1103.
吴颖谦等.基于小波变换的低对比度图像增强[J].红外与激光工程,2003.32(2):4 6.
李涛，小波理论及其在图像处理中的应用：[硕士学位论文].西安：西安电子科技大学.
许雷,郑筱祥,陈兴灿等.一种基于小波相位滤波及视觉非线性的医学图像自适应增强新方法[Ｊ].电子学报,1999,27(9):121 123.