基于图像处理的鉴定身份证真伪技术的研究与实现
|
摘要
身份证是居民身份的象征,它与人们的日常生活息息相关。近几年来,伪造身份证犯罪活动日益猖獗,使国家一些重要部门遭受到严重的损失,如何快速准确地辨认身份证的真伪,是急需解决的问题。
本文研究如何运用图像处理的知识,利用计算机自动鉴别身份证真伪的技术。文中详细阐述了整个鉴别身份证真伪系统的处理流程,比较完整地介绍了该系统的预处理、定位、图像分割、特征提取、相似性度量等功能模块,并用MATLAB软件编写了一个智能鉴别身份证真伪的系统。
通过对身份证图像的色彩分析,本文提出了特殊的预处理方法,利用R、G、B分量之间的差值对身份证图像进行色彩转换,并利用整体阈值法把图像二值化,用数学形态学对二值图像进行滤波。这种预处理方法解决了如何在具有复杂背景与强烈干扰的身份证图像中搜索目标的问题。
本文以激光防伪标识为研究日标,运用投影法对它进行定位,并针对其特点,从分块方式和阈值调整两方面对传统的局部阈值法进行了改进,把这种改进后的方法用于防伪标识的图像分割,可以得到清晰的目标。
本文对如何提取目标的形状特征进行了研究,并提出了一种新的方法——区域方向直方图来表达区域内各点分布的情况,这种直方图由角向直方图和径向直方图共同组成,考虑了目标区域内各点的角向和径向分布,能够全面、准确地表达图像的形状特征。用区域方向直方图抽取目标的形状特征,判断是否存在清晰的激光防伪标志,从而鉴别身份证的真伪。
本文选取了2123张真身份证、90张假身份证进行测试,正确率达到99.46%,实验证明,本文提出的智能鉴别身份证真伪系统达到实用要求,解决了利用机器智能识别证件真伪的问题,增强了证件防伪的效果,提高了工作效率,能有效地打击伪造身份证的现象。
ID card is an identity symbol of residents. It is closely related to our lives. However, crime of forging ID card is so frantic that a number of departments in our country have suffered severe loss. So, it is urgent to distinguish true-false of ID card quickly and exactly.Distinguishing true-false of ID card based on image processing is researched in this paper. It elaborated the processing flow of the distinguishing true-false of ID card system. Function modules, such as pretreatment, localization, image segmentation, extraction of image feature and so on, were introduced entirely. And we compiled a distinguishing true-false of ID card system with MATLAB software.Through color analysis, one special method of pretreatment for the ID card image was proposed. It was composed by color conversion using difference between R, G, B component, binarization using whole threshold method, and filtering with mathematical morphology. This pretreatment method solved the problem on how to search aim in the ID card image which has complex background and intensive interaction.Laser identity logo was the research aim in this paper. It was localized by projection method. According to the identity logo's characteristics, we improved the traditional local-threshold method from two aspects which included blocking ways and threshold adjustment. We used it for image segmentation and we can obtain a sharp image.In this paper, we made an intensive study of extraction of shape feature. A new method which named region direction histogram was proposed. This kind of histogram was formed from angle histogram and radial histogram. It considers angle and radial distributing of each point in the area. So it can express image's
本文研究如何运用图像处理的知识,利用计算机自动鉴别身份证真伪的技术。文中详细阐述了整个鉴别身份证真伪系统的处理流程,比较完整地介绍了该系统的预处理、定位、图像分割、特征提取、相似性度量等功能模块,并用MATLAB软件编写了一个智能鉴别身份证真伪的系统。
通过对身份证图像的色彩分析,本文提出了特殊的预处理方法,利用R、G、B分量之间的差值对身份证图像进行色彩转换,并利用整体阈值法把图像二值化,用数学形态学对二值图像进行滤波。这种预处理方法解决了如何在具有复杂背景与强烈干扰的身份证图像中搜索目标的问题。
本文以激光防伪标识为研究日标,运用投影法对它进行定位,并针对其特点,从分块方式和阈值调整两方面对传统的局部阈值法进行了改进,把这种改进后的方法用于防伪标识的图像分割,可以得到清晰的目标。
本文对如何提取目标的形状特征进行了研究,并提出了一种新的方法——区域方向直方图来表达区域内各点分布的情况,这种直方图由角向直方图和径向直方图共同组成,考虑了目标区域内各点的角向和径向分布,能够全面、准确地表达图像的形状特征。用区域方向直方图抽取目标的形状特征,判断是否存在清晰的激光防伪标志,从而鉴别身份证的真伪。
本文选取了2123张真身份证、90张假身份证进行测试,正确率达到99.46%,实验证明,本文提出的智能鉴别身份证真伪系统达到实用要求,解决了利用机器智能识别证件真伪的问题,增强了证件防伪的效果,提高了工作效率,能有效地打击伪造身份证的现象。
ID card is an identity symbol of residents. It is closely related to our lives. However, crime of forging ID card is so frantic that a number of departments in our country have suffered severe loss. So, it is urgent to distinguish true-false of ID card quickly and exactly.Distinguishing true-false of ID card based on image processing is researched in this paper. It elaborated the processing flow of the distinguishing true-false of ID card system. Function modules, such as pretreatment, localization, image segmentation, extraction of image feature and so on, were introduced entirely. And we compiled a distinguishing true-false of ID card system with MATLAB software.Through color analysis, one special method of pretreatment for the ID card image was proposed. It was composed by color conversion using difference between R, G, B component, binarization using whole threshold method, and filtering with mathematical morphology. This pretreatment method solved the problem on how to search aim in the ID card image which has complex background and intensive interaction.Laser identity logo was the research aim in this paper. It was localized by projection method. According to the identity logo's characteristics, we improved the traditional local-threshold method from two aspects which included blocking ways and threshold adjustment. We used it for image segmentation and we can obtain a sharp image.In this paper, we made an intensive study of extraction of shape feature. A new method which named region direction histogram was proposed. This kind of histogram was formed from angle histogram and radial histogram. It considers angle and radial distributing of each point in the area. So it can express image's
引文
[1] 荣海泓,席志红.生物特征身份鉴别技术应用及发展趋势[J].现代电子技术,2004,10:88～91
[2] Shen W, Marc S, Rajiv K. Evalution of automated biometrics-based identification and verification systems [J]. Proceeding of The IEEE,1997,85(9):1464～1478
[3] 张敏贵,周德龙,潘泉等.生物特征识别及研究现状.生物物理学报,2002,18(2):156～162
[4] Woodward JD. Biometrics: privacy's foe or privacy's friend? [J]. Proceeding of The IEEE,1997,85(9): 1480～1492
[5] 高康林.金融、电子商务中的生物安全识别技术及应用[J].现代电子技术,2004,1:73～76
[6] 徐飞,施晓红等.MATLAB应用图像处理[M].西安:西安电子科技大学出版社,2002.17-18
[7] 沈瑛.身份证识别系统研究.西南交通大学硕士学位论文,2002
[8] 涂其远,吴建华,万国金.动态阈值结合全局阈值对图像进行分割[J].南昌大学学报(工科版),2002,24(1):37～40
[9] 付忠良.一些新的图像阈值选取方法[J].计算机应用,2000,20(10):13～15
[10] 韩思奇,王蕾.图像分割的阈值法综述[J].系统工程与电子技术,2002,24(6):91～94
[11] 吴冰,秦志远.自动确定图像二值化最佳阈值的新方法[J].测绘学院学报,2001,18(4):283～286
[12] Doyle W. Operation Useful for Similarity - Invariant Pattern Recognition [J]. J. Asssoc. Comput. Mach. ,1962,9:259～267
[13] 程宏煌,戴卫恒,姚梗梗.图像分割方法综述[J].电信快报,2000,10:39～41
[14] Sahoo P K. A survey of Threshold technigues, Comput.Vision Graphics, Image Process. 1988,41:233～260
[15] 付忠良.图像阈值选取方法——Otsu方法的推广[J].计算机应用,2000, 20(5): 37～39
[16] Ostu N A. Threshold Selection Method from Gray-Level Histograms [J]. IEEE Trans. on System,Man, and Cybemetics, 1979,9(1):62～66
[17] 杨述斌.数学形态学在图像滤波中的应用[J].褒樊学院学报,2001,22(2):33～35
[18] 章毓晋.图象处理和分析[M].北京:清华大学出版社,2001.82～98
[19] 郭景峰,申光宪,郑绳楦等.数学形态学在数字滤波中的应用研究[J].机械工程学报,2002,38(10):144～147
[20] 王丰元,高松,孙壮志.数学形态学原理及其在图象处理中的应用[J].山东工程学院学报,1999,13(2):5～8
[21] Maragos P. Tutorial on advances in morphological image processing an analysis [J]. Optical Engineering, 1987, 26(7): 623～632
[22] 马然,陈颖.图象处理中数学形态学方法的基本研究[J].上海轻工业高等专科学校学报,2000,21(1):17～20
[23] 戴青云.数学形态学在图像处理中的应用进展[J].控制理论与应用,2001,18(4):478～482
[24] 张怀朋.身份号码图像识别方法与系统的研究.大学硕士学位论文,2003
[25] 廖金周,宣国荣.车辆牌照的自动分割[J].微型电脑应用研究与设计,1997,7:32～37
[26] 张引,潘云鹤.面向车辆牌照字符识别的预处理算法[J].计算机研究与应用,1999.7:85～87
[27] Castleman K.R.. Digital Image Processing. New Jersey: Prentice Hall,1996
[28] Pal N.R. , Pal S.K.. A review on image segmentation techniques [J]. Pattern Recognition, 1993,26(9): 1277～1294
[29] 赵荣椿,迟耀斌,朱重光.图像分割技术进展[J].中国体视学与图像分析,1998,3(2):121～128
[30] 刘秀兰,马丹,刘兵等.一种新的快速动态阈值图像分割算法[J].北京工业大学学报,1999,25(2):92～96
[31] 贺霖,李见为.基于零交叉检测器的动态阈值选取[J].光电工程,2003,30(4):65～68
[2] Shen W, Marc S, Rajiv K. Evalution of automated biometrics-based identification and verification systems [J]. Proceeding of The IEEE,1997,85(9):1464～1478
[3] 张敏贵,周德龙,潘泉等.生物特征识别及研究现状.生物物理学报,2002,18(2):156～162
[4] Woodward JD. Biometrics: privacy's foe or privacy's friend? [J]. Proceeding of The IEEE,1997,85(9): 1480～1492
[5] 高康林.金融、电子商务中的生物安全识别技术及应用[J].现代电子技术,2004,1:73～76
[6] 徐飞,施晓红等.MATLAB应用图像处理[M].西安:西安电子科技大学出版社,2002.17-18
[7] 沈瑛.身份证识别系统研究.西南交通大学硕士学位论文,2002
[8] 涂其远,吴建华,万国金.动态阈值结合全局阈值对图像进行分割[J].南昌大学学报(工科版),2002,24(1):37～40
[9] 付忠良.一些新的图像阈值选取方法[J].计算机应用,2000,20(10):13～15
[10] 韩思奇,王蕾.图像分割的阈值法综述[J].系统工程与电子技术,2002,24(6):91～94
[11] 吴冰,秦志远.自动确定图像二值化最佳阈值的新方法[J].测绘学院学报,2001,18(4):283～286
[12] Doyle W. Operation Useful for Similarity - Invariant Pattern Recognition [J]. J. Asssoc. Comput. Mach. ,1962,9:259～267
[13] 程宏煌,戴卫恒,姚梗梗.图像分割方法综述[J].电信快报,2000,10:39～41
[14] Sahoo P K. A survey of Threshold technigues, Comput.Vision Graphics, Image Process. 1988,41:233～260
[15] 付忠良.图像阈值选取方法——Otsu方法的推广[J].计算机应用,2000, 20(5): 37～39
[16] Ostu N A. Threshold Selection Method from Gray-Level Histograms [J]. IEEE Trans. on System,Man, and Cybemetics, 1979,9(1):62～66
[17] 杨述斌.数学形态学在图像滤波中的应用[J].褒樊学院学报,2001,22(2):33～35
[18] 章毓晋.图象处理和分析[M].北京:清华大学出版社,2001.82～98
[19] 郭景峰,申光宪,郑绳楦等.数学形态学在数字滤波中的应用研究[J].机械工程学报,2002,38(10):144～147
[20] 王丰元,高松,孙壮志.数学形态学原理及其在图象处理中的应用[J].山东工程学院学报,1999,13(2):5～8
[21] Maragos P. Tutorial on advances in morphological image processing an analysis [J]. Optical Engineering, 1987, 26(7): 623～632
[22] 马然,陈颖.图象处理中数学形态学方法的基本研究[J].上海轻工业高等专科学校学报,2000,21(1):17～20
[23] 戴青云.数学形态学在图像处理中的应用进展[J].控制理论与应用,2001,18(4):478～482
[24] 张怀朋.身份号码图像识别方法与系统的研究.大学硕士学位论文,2003
[25] 廖金周,宣国荣.车辆牌照的自动分割[J].微型电脑应用研究与设计,1997,7:32～37
[26] 张引,潘云鹤.面向车辆牌照字符识别的预处理算法[J].计算机研究与应用,1999.7:85～87
[27] Castleman K.R.. Digital Image Processing. New Jersey: Prentice Hall,1996
[28] Pal N.R. , Pal S.K.. A review on image segmentation techniques [J]. Pattern Recognition, 1993,26(9): 1277～1294
[29] 赵荣椿,迟耀斌,朱重光.图像分割技术进展[J].中国体视学与图像分析,1998,3(2):121～128
[30] 刘秀兰,马丹,刘兵等.一种新的快速动态阈值图像分割算法[J].北京工业大学学报,1999,25(2):92～96
[31] 贺霖,李见为.基于零交叉检测器的动态阈值选取[J].光电工程,2003,30(4):65～68