体全息子波相关器及其旋转不变性研究
|
摘要
随着社会信息量“爆炸”似的增长,原来的存储器件和存取速度越来越不够用。体全息存储适时而生,它有着高冗余度、高存储容量、极高的数据传输速率以及极快的存取时间等优点,因而得到越来越多的关注。本论文以体全息相关器为基础进行图像识别。
本论文介绍了体全息存储原理及技术,包括体光栅的形成和布拉格衍射、衍射效率、水平角度选择性,还介绍了相干光学信息处理系统、Vander Lugt 滤波器的形成过程,给出了多重匹配滤波器相关的运算过程,最后介绍了子波变换的定义以及光学子波变换。
本论文构建了一个与北京市车辆牌照密切相关的牌照图像库。从仿真和实验两方面比较了空域滤波与频域滤波、一般高通滤波器与子波滤波器的优缺点。实验表明子波滤波器的引入大大提高了相关器的识别能力。
为了解决旋转不变识别,本论文采用综合判别函数方法,对每一个畸变训练集,综合出一个综合判别函数。在光路引入子波滤波器后,利用角度复用技术将综合判别函数依次存入晶体,然后输入任意一个训练集的任意一幅图像,都被精确地识别。
本论文尝试着采用圆谐展开方法来实现旋转不变识别。选取一幅图像的圆谐二级分量,将其生成纯位相滤波器,然后存入晶体。输入任意旋转角度的图像,都被准确地识别出来。为了减少滤波器的个数和计算量,结合圆谐展开和综合判别函数方法生成圆谐—综合判别函数滤波器。利用该滤波器可以识别多类任意旋转角度的图像。
本论文还研究了三维物体相关识别,发现三维物体的相关输出对旋转角度极为敏感,这一特点可以应用于机器人视觉识别。
本论文最后总结了取得的成果,并指明了以后的研究方向。
Due to the “information explosion”, Memory capacity and its access speed have not satisfied our need in the last decade. More attention are attached on volume holographic storage for its high redundancy, capacity, fast data transfer rate and short access time. Image recognition is carried out based on volume holographic storage.
In this paper, principle and technique of volume holographic storage are presented such as volume grating formation, Bragg diffraction, and horizontal angle selectivity. Also optical information processing system and Vander Lugt filter are presented, as well as correlation procedure based on matched-filter. Finally optical wavelet transform is presented.
In this paper, an image set is constructed according to registration mark in Beijing District. From computer simulation and experiments comparison results are presented between spatial and frequency , high-pass and wavelet filtering. And it is suggested that the discrimination is greatly improved owing to the wavelet filter.
To achieve rotation-invariance, Synthetic Discriminant Function(SDF) is introduced . A SDF is synthesized using a
本论文介绍了体全息存储原理及技术,包括体光栅的形成和布拉格衍射、衍射效率、水平角度选择性,还介绍了相干光学信息处理系统、Vander Lugt 滤波器的形成过程,给出了多重匹配滤波器相关的运算过程,最后介绍了子波变换的定义以及光学子波变换。
本论文构建了一个与北京市车辆牌照密切相关的牌照图像库。从仿真和实验两方面比较了空域滤波与频域滤波、一般高通滤波器与子波滤波器的优缺点。实验表明子波滤波器的引入大大提高了相关器的识别能力。
为了解决旋转不变识别,本论文采用综合判别函数方法,对每一个畸变训练集,综合出一个综合判别函数。在光路引入子波滤波器后,利用角度复用技术将综合判别函数依次存入晶体,然后输入任意一个训练集的任意一幅图像,都被精确地识别。
本论文尝试着采用圆谐展开方法来实现旋转不变识别。选取一幅图像的圆谐二级分量,将其生成纯位相滤波器,然后存入晶体。输入任意旋转角度的图像,都被准确地识别出来。为了减少滤波器的个数和计算量,结合圆谐展开和综合判别函数方法生成圆谐—综合判别函数滤波器。利用该滤波器可以识别多类任意旋转角度的图像。
本论文还研究了三维物体相关识别,发现三维物体的相关输出对旋转角度极为敏感,这一特点可以应用于机器人视觉识别。
本论文最后总结了取得的成果,并指明了以后的研究方向。
Due to the “information explosion”, Memory capacity and its access speed have not satisfied our need in the last decade. More attention are attached on volume holographic storage for its high redundancy, capacity, fast data transfer rate and short access time. Image recognition is carried out based on volume holographic storage.
In this paper, principle and technique of volume holographic storage are presented such as volume grating formation, Bragg diffraction, and horizontal angle selectivity. Also optical information processing system and Vander Lugt filter are presented, as well as correlation procedure based on matched-filter. Finally optical wavelet transform is presented.
In this paper, an image set is constructed according to registration mark in Beijing District. From computer simulation and experiments comparison results are presented between spatial and frequency , high-pass and wavelet filtering. And it is suggested that the discrimination is greatly improved owing to the wavelet filter.
To achieve rotation-invariance, Synthetic Discriminant Function(SDF) is introduced . A SDF is synthesized using a
引文
1.Fai H.Mok and Harold M.Stoll, Holographic inner-processor for pattern recognition, SPIE, Vol.1701, Optical Pattern Recognition III(1992)
2.Allen Pu, et al. Real-time vehicle navigation using a holographic memory, Opt.Eng, 1997, Vol.18, No.11, 915-917
3.Hua-Kuang Liu, Yahong Jin, and Nevillle I.Marzwell, Advanced ultra-high-capacity optical random access memory and pattern recognition techniques, Opt.Eng, Vol.37, No.3, 1998,119-788
4.孙骞等,光折变光学相关器在防伪中的应用,光电子﹒激光,1996,Vol.7, No.5,295-297
5. 冯文毅等,光折变晶体体全息子波相关图像识别系统,光电子.激光,2000, Vol.11,No.2,133-136
6. 冯文毅等,多通道体全息子波相关器及其在指纹识别中的应用,光学学报, 2000,Vol.20,No.3,370-375
7. 冯文毅等,体全息子波相关系统并行性的研究,光子学报,1999,Vol.28, No.5,414-419
8. 周雁,采用体全息图像库的模式识别,北京工业大学硕士论文,2000
9. 陈孙征,用体全息相关器和多重频谱滤波法做图象识别,北京工业大学硕士 论文,2002
10. FRANCIS T.S.YU, Optical Pattern Recognition: Architectures and Techniques, proceedings of the IEEEE, 1996, Vol.84, No.5, 733-752
11.吴华文等,光学小波匹配滤波器与图形识别,物理,2002,No.10,654-658
12.Yunlong Sheng and Danny Roberge, Optical wavelet matched filter for shift-invariant pattern recognition, Optics Letters, 1993,Vol.18, No.4, 299-301
13.R.K.Wang, et al. Modified filter synthetic discriminant functions for improved optical correlator performance, App.Opt, 1994, Vol.33, No.32, 7646-7654
14.David Casasent, Unified synthetic discriminant function computational formulation, App.Opt, 1984, Vol.23, No.10, 1620-1627
15.Zikuan Chen, et al. Pattern identification with an improved synthetic discrimination function, Opt.Eng, 1994, Vol.33, No.9, 3044-3046
16.Vahid R.Riasati, Rotation-invariant synthetic discriminant function filter for pattern recognition, Opt.Eng, Vol.39, No.5, 1156-1161
17.Yang-hoi Doh, Distortion-invariant pattern recognition based on a synthetic hit-miss transform, Opt.Eng, Vol.43, No.8, 1798-1803
18.何万涛等,光电混合综合判别相关器,光学学报,1996,Vol.16, No.1, 100-105
19.陈前荣等,用等相关峰法或迭代法合成旋转不变综合判别函数,光电子.激 光,2003,Vol.14, No.3, 299-302
20.何万涛等,综合判别函数相关器实验研究,激光技术,1996,Vol.20, No.2,122-125
21. 孙红辉等,给予计算机处理的畸变图像相关识别,机电工程技术,2004, Vol.33, No.6, 44-46
22. Yuan-Neng Hsu and H.H.Arsenault, Optical pattern recognition using circular harmonic expansion, App.Opt, 1982, Vol.21, No.22, 4016-4019
23.Yunlong Sheng and Henri H.Arsenault, Method for determining expansion centers and predicting sidelobe levels for circular-harmonic filters, J.Opt.Soc.Am.A, 1987, Vol.4, No.9, 1793-1797
24.Y.N.Hsu, et al. Rotation-invariant digital pattern recognition using circular harmonic expansion, App.Opt, 1982, Vol.21, 4012-4015
25. H.H.Arsennault, et al. Rotation-invariant pattern recognition, Opt.Eng, 1984, Vol.23, 705-709
26. Luc Leclerc, et al. Optical binary phase-only filters for circular harmonic correlations, App.Opt, Vol.30, No.32, 4643-4649
27. Hon Fai Yau and Chi Ching Chang, Phase-only circular harmonic matched filtering, App.Opt, Vol.28, No.11, 2070-2074
28. Hon Fai Yau, et al. Experimental demonstration of phase-only circular harmonic filtering using computer generated filters, App.Opt, Vol.29, No.23, 3387-3391
29. 王培锋等,基于二元纯位相匹配滤波器的光电混合模式识别,红外和毫米波 学报,2000,Vol.19, No.2, 117-120
30. 陈学文等,旋转不变振幅位相合成圆谐滤波器,光学学报,1996,Vol.16, No.3, 373-377
31.黄晓菁,用二元小波滤波器作图像相关识别,光电子﹒激光,2001,Vol.12, No.10, 1068-1071
32.Dan Cojoc, Rotation invariant pattern recognition using circular harmonics and synthetic discriminant functions, SPIE, 1998, Vol.3405, 425-431
33.Seung-Ho Shin, et al. Three-dimensional object recognition by use of a photorefractive volume holographic processor, Opt.Lett, 2001, Vol.26, No.15, 1161-1163
34 . B.V.K.Vijaya Kumar and L.Hassebrook, Performance measures for correlation filters, App.Opt, 1990, Vol.29, No.20, 2997-3000
35.陶世荃等,光全息存储,北京工业大学出版社,1998
36.J.W.Goodman,傅里叶光学导论,科学出版社,1975
37. 金国藩等,光学子波变换(1)基本理论,光电子﹒激光,1994,Vol.5,No.4, 193-212
38. 王文陆等,光学子波变换(2)基本理论,光电子﹒激光,1994,Vol.5,No.5, 268-272
39. 王文陆等,光学子波变换(3)基本理论,光电子﹒激光,1994,Vol.5,No.6, 212-339
40.庞绍芳等,光学小波变换的伸缩因子极限,辽宁大学学报,2003,Vol.30, No.4,332-334
41.吕乃光,傅里叶光学,机械工业出版社,1998
2.Allen Pu, et al. Real-time vehicle navigation using a holographic memory, Opt.Eng, 1997, Vol.18, No.11, 915-917
3.Hua-Kuang Liu, Yahong Jin, and Nevillle I.Marzwell, Advanced ultra-high-capacity optical random access memory and pattern recognition techniques, Opt.Eng, Vol.37, No.3, 1998,119-788
4.孙骞等,光折变光学相关器在防伪中的应用,光电子﹒激光,1996,Vol.7, No.5,295-297
5. 冯文毅等,光折变晶体体全息子波相关图像识别系统,光电子.激光,2000, Vol.11,No.2,133-136
6. 冯文毅等,多通道体全息子波相关器及其在指纹识别中的应用,光学学报, 2000,Vol.20,No.3,370-375
7. 冯文毅等,体全息子波相关系统并行性的研究,光子学报,1999,Vol.28, No.5,414-419
8. 周雁,采用体全息图像库的模式识别,北京工业大学硕士论文,2000
9. 陈孙征,用体全息相关器和多重频谱滤波法做图象识别,北京工业大学硕士 论文,2002
10. FRANCIS T.S.YU, Optical Pattern Recognition: Architectures and Techniques, proceedings of the IEEEE, 1996, Vol.84, No.5, 733-752
11.吴华文等,光学小波匹配滤波器与图形识别,物理,2002,No.10,654-658
12.Yunlong Sheng and Danny Roberge, Optical wavelet matched filter for shift-invariant pattern recognition, Optics Letters, 1993,Vol.18, No.4, 299-301
13.R.K.Wang, et al. Modified filter synthetic discriminant functions for improved optical correlator performance, App.Opt, 1994, Vol.33, No.32, 7646-7654
14.David Casasent, Unified synthetic discriminant function computational formulation, App.Opt, 1984, Vol.23, No.10, 1620-1627
15.Zikuan Chen, et al. Pattern identification with an improved synthetic discrimination function, Opt.Eng, 1994, Vol.33, No.9, 3044-3046
16.Vahid R.Riasati, Rotation-invariant synthetic discriminant function filter for pattern recognition, Opt.Eng, Vol.39, No.5, 1156-1161
17.Yang-hoi Doh, Distortion-invariant pattern recognition based on a synthetic hit-miss transform, Opt.Eng, Vol.43, No.8, 1798-1803
18.何万涛等,光电混合综合判别相关器,光学学报,1996,Vol.16, No.1, 100-105
19.陈前荣等,用等相关峰法或迭代法合成旋转不变综合判别函数,光电子.激 光,2003,Vol.14, No.3, 299-302
20.何万涛等,综合判别函数相关器实验研究,激光技术,1996,Vol.20, No.2,122-125
21. 孙红辉等,给予计算机处理的畸变图像相关识别,机电工程技术,2004, Vol.33, No.6, 44-46
22. Yuan-Neng Hsu and H.H.Arsenault, Optical pattern recognition using circular harmonic expansion, App.Opt, 1982, Vol.21, No.22, 4016-4019
23.Yunlong Sheng and Henri H.Arsenault, Method for determining expansion centers and predicting sidelobe levels for circular-harmonic filters, J.Opt.Soc.Am.A, 1987, Vol.4, No.9, 1793-1797
24.Y.N.Hsu, et al. Rotation-invariant digital pattern recognition using circular harmonic expansion, App.Opt, 1982, Vol.21, 4012-4015
25. H.H.Arsennault, et al. Rotation-invariant pattern recognition, Opt.Eng, 1984, Vol.23, 705-709
26. Luc Leclerc, et al. Optical binary phase-only filters for circular harmonic correlations, App.Opt, Vol.30, No.32, 4643-4649
27. Hon Fai Yau and Chi Ching Chang, Phase-only circular harmonic matched filtering, App.Opt, Vol.28, No.11, 2070-2074
28. Hon Fai Yau, et al. Experimental demonstration of phase-only circular harmonic filtering using computer generated filters, App.Opt, Vol.29, No.23, 3387-3391
29. 王培锋等,基于二元纯位相匹配滤波器的光电混合模式识别,红外和毫米波 学报,2000,Vol.19, No.2, 117-120
30. 陈学文等,旋转不变振幅位相合成圆谐滤波器,光学学报,1996,Vol.16, No.3, 373-377
31.黄晓菁,用二元小波滤波器作图像相关识别,光电子﹒激光,2001,Vol.12, No.10, 1068-1071
32.Dan Cojoc, Rotation invariant pattern recognition using circular harmonics and synthetic discriminant functions, SPIE, 1998, Vol.3405, 425-431
33.Seung-Ho Shin, et al. Three-dimensional object recognition by use of a photorefractive volume holographic processor, Opt.Lett, 2001, Vol.26, No.15, 1161-1163
34 . B.V.K.Vijaya Kumar and L.Hassebrook, Performance measures for correlation filters, App.Opt, 1990, Vol.29, No.20, 2997-3000
35.陶世荃等,光全息存储,北京工业大学出版社,1998
36.J.W.Goodman,傅里叶光学导论,科学出版社,1975
37. 金国藩等,光学子波变换(1)基本理论,光电子﹒激光,1994,Vol.5,No.4, 193-212
38. 王文陆等,光学子波变换(2)基本理论,光电子﹒激光,1994,Vol.5,No.5, 268-272
39. 王文陆等,光学子波变换(3)基本理论,光电子﹒激光,1994,Vol.5,No.6, 212-339
40.庞绍芳等,光学小波变换的伸缩因子极限,辽宁大学学报,2003,Vol.30, No.4,332-334
41.吕乃光,傅里叶光学,机械工业出版社,1998