臺灣博碩士論文加值系統

即時影像識別在光學資訊處理中是很重要的一項。其中，聯合轉換相關器對於光學影像圖形的辨認方面極為有用。此外，再配合近代空間光調制器的技術後，使得聯合轉換相關器變得穩定而且容易實現。
然而，傳統的聯合轉換相關器卻存在著一些問題。主要問題之一是零階的直流項能量過於強烈，因為這個直流項是由參考影像和輸入影像彼此自相關的總和。也因此，直流項常常覆蓋到我們所須要偵測的訊號。所以，對於參考影像和輸入影像之間的距離也總是被限制著。
所謂的二元化是對平面作二元化非線性處理,使得傅立葉頻譜二元化。這個技術包括固定橫切閥線，適應性閥線，閥線函數，中間值和區域中間值等方法。有人指出，與傳統的聯合轉換相關器比較時，二元化聯合轉換相關器在單物件辨認時可以提供更高的相關項強度，更狹窄的相關項寬度和更好的辨識性。但是，二元化聯合轉換相關器卻在輸出面上產生出較高的諧波項，這將增大判斷時誤判的可能性。因此，最近提出了餘弦波編碼聯合轉換相關器。這種方法能夠有效的解決那些諧波項的問題。但是，其裝置卻是很複雜難以實現。
在這本論文中，我們介紹適應性平均聯合轉換相關器技術，這是使用適應性平均值方程式來修改傅立葉頻譜圖。這方法能夠使得傅立葉頻譜正常化並且能有效地取出餘弦波的特性。所以，與傳統的聯合轉換相關器比較時，適應性平均聯合轉換在各式物件辨認時可以提供更高的相關項強度，更狹窄的相關項寬度和更好的辨識性。更重要的特性是適應性平均聯合轉換相關器適應能夠有效的解決那些諧波項的問題，而且裝置卻是很簡單並且容易實現。

For the real-time image recognition, the optical information processing is very important. In the optical information processing, the optical joint transform correlator (JTC) is extremely useful for optical pattern recognition. Moreover, the JTC is very stable and easy to set up with modern spatial light modulator (SLM) technology.
However, there are several difficulties associated with the conventional JTC. One of the main problems is the presence of a strong zero-order peak in the output plane that corresponds to the sum of the autocorrelations of the reference and input images, and always overlaps the desired correlation signals. So, the distance between the reference and input images is always to be restricted.
The binary JTC (BJTC) uses binary nonlinearity procession in the Fourier plane to binarize the joint power spectrum (JPS) of the reference and input images. This technique includes hard-clipping thresholding, adaptive thresholding, threshold functions, and median and submedian thresholding methods. It has been shown that, compared with the classical JTC (CJTC), the BJTC provides a higher correlation peak intensity, narrower peak width, and better discrimination sensitivity for similar objects. Even though this, the BJTC also induces higher harmonic terms in the output plane, which may cause false alarms. So, recently a cosine wave encoded JTC (CWJTC) has been proposed. The proposed method can be effective to solve the effect of the harmonic terms. But, the system is very complex and not easy to realize.
In this paper, we introduce the adaptive mean JTC (AMJTC) technique that uses the adaptive mean function to modify the JPS. The proposed the AMJTC can normalize the JPS and reserve the effective cosine wave features. And, it has been shown that, compared with the CJTC, the AMJTC provides a higher correlation peak intensity, narrower peak width, and better discrimination sensitivity for similar objects and multiple objects. The important characteristics are that the AMJTC can be effective to solve the problems of harmonic terms and it is very easy to realize the system.

ABSTRACT IN CHINESE
ABSTRACT IN ENGLISH
ACKNOWLEDGMENTS
CONTENTS
TABLE OF CONTENTS
LIST OF FIGURES
Chapter 1 INTRODUCTION
1.1 Purpose
1.2 Thesis organization
Chapter 2 BASIC THEORY
2.1 Fourier analysis in two dimensions
2.1.1 Definition and existence conditions
2.1.2 Fourier transform theorem
2.1.3 Separable functions
2.1.4 Some frequently used functions and Fourier transform
2.2 Fresnel and Fraunhofer Principle
2.2.1 Approximations of the Huygens-Fresnel principle
2.2.2 Initial approximations
2.2.3 The Fresnel approximations
2.3 A thin lens as a phase transformation
2.3.1 Fourier transforming properties of lenses
2.3.2 Object leaned against the lens
2.3.3 Object leaned in front of the lens
Chapter 3 JOINT TRANSFORM CORRELATOR
3.1 Summary of the spatial filter
3.1.1 Simulation for the spatial filter
3.2 Classical joint transform correlator
3.2.1 Analysis
3.2.2 Simulation
3.2.3 Discussion
3.3 Binary joint transform correlator
3.3.1 Analysis
3.3.2 Simulation
3.3.3 Discussion
3.4 Cosine wave encoded joint transform correlation
3.4.1 Analysis
3.4.2 Simulation
3.4.3 Discussion
Chapter 4 PROPOSED ADAPTIVE JOINT TRANSFORM CORRELATOR
4.1 Introduction
4.1.1 AMJTC
4.2 Analysis
4.2.1. Adaptive mean processing using the target and the reference located at the identical axis
4.2.2. Adaptive mean processing using the target and the reference not located at the identical axis
4.3 Comparison of various JTCs using one single target
4.3.1. Comparison of various JTCs using one single target without noise
4.3.2. Comparison of various JTCs using one single target with nonoverlapping noise
4.4 Simulation of various JTCs using multiple objects
4.4.1. N targets located at different positions
4.4.2. JTC using three targets located at identical axis
4.4.3. JTC using N targets located at identical axis
Chapter 5 EXPERIMENTAL RESULTS
5.1 System description
5.2 Experimental Results
Chapter 6 CONCLUSIONS
REFERENCES
VITA
PUBLICATION LIST OF JUN-DA LIN

[1]J. L. Horner and J. R. Leger, “Pattern recognition with binary phase-only filters,” Appl. Opt. 24, P. 609 (1985).
[2]J. L. Horner and H. O. Bartelt, “Two-bit correlator,” Appl. Opt. 24, P. 2889 (1985).
[3]B. Javidi, “Nonlinear joint power spectrum based optical correlator,” Appl. Opt. 28, P. 2358 (1989).
[4]K. H. Fielding and J. L. Horner, “1-f binary joint transform correlation,” Opt. Eng. 29, P. 1081 (1990).
[5]S. K. Rogers, J. D. Cline and M. Kabrisky, “New binarization techniques for joint transform correlation,” Opt. Eng. 29, P. 1088 (1990).
[6]C. S. Weaver and J. W. Goodman, “A technique for optically convolving two functions,” Appl. Opt. 5, P. 1248 (1966).
[7]F. Lei, H. Huang, N. Yoshikawa, M. Iton and T. Yatagai, “Performance of binary joint transform correlator using linear combination theshold function,” Opt. Commun. 169, P. 207 (1999).
[8]F. T. S. Yu, F. Cheng, T. Nagata and D. A. Gregory, ”Effects of fringe binarization of multiobject joint transform correlation,” Appl. Opt. 28, P. 2988 (1989).
[9]B. Javidi, J. Li, A. H. Fazlollahi and Joseph Horner, ”Binary nonlinear joint transform correlator performance with different thresholding methods under unknown illumination conditions,” Appl. Opt. 34, P. 886 (1995).
[10]G. S. Patil and K. Singh, “Experimental and simulation studies on performance of binary and gray-valued joint transform correlator under poor illumination conditions and nonoverlapping background noise,” Opt. Eng. 36, P. 1918 (1997).
[11]W. B. Hahn and D. L. Flannery, “Design elements of a binary joint transform correlator and selected optimization techniques,” Opt. Eng. 31, P. 896 (1992).
[12]M. S. Alam and M. A. Karim, “Fringe-adjusted joint transform correlation,” Appl. Opt. 32, P. 4344 (1993).
[13]R. K. Wang, L. Shang and C. R. chatwin, “Modified fringe-adjusted joint transform correlation to accommodate noise in the input scene,” Appl. Opt. 35, P. 286 (1996).
[14]R. K. Wang, L. Shang, and C. R. Chatwin, “Cosine wave encoded joint transform correlation,” Appl. Opt. 35, P. 296 (1996).
[15]M. S. Alam, “Phase-encoded fringe-adjusted joint transform correlation,” Opt. Eng. 39, p. 1169 (2000).
[16]D. A. Gregory, “Real-time pattern recognition using a modified liquid crystal television in a coherent optical correlator,” Appl. Opt. 25, P. 467 (1986).
[17]F. T. S. Yu, S, Jutamulia, T. W. Lin and D. A. Gregory, “Adaptive real-time pattern recognition using a liquid crystal TV based joint transform correlator,” Appl. Opt. 26, P. 1370 (1987).
[18]B. Javidi, D. A. Gregory and J. L. Horner, “Single modulator joint transform correlator architectures,” Appl. Opt. 28, P. 411 (1989).
[19]B. Javidi and J. L. Horner, “Single SLM joint transform correlator,” Appl. Opt. 28, P. 1027 (1989).
[20]F. T. S. Yu, Q. W. Song, Y. S. Cheng and D. A. Gregory, “Comparison of detection efficiencies for VanderLugt and Joint transform correlators,” Appl. Opt. 29, P. 225 (1990).
[21]Bracewell, R. N., “ The fourier transform and its application,” McGraw-Hill Book Company, New York (1965).
[22]Eugene Hecht, ”Optics,” 3rd, Addison Wesley longman, Inc (1998).
[23]J. W. Goodman, “Introduction to fourier optics,” McGraw-Hill Book Company, Inc (1998).
[24]F. T. S. Yu, Chun-Te Li and S. Yin, “Comparison of detection efficiencies for VanderLugt and joint transform correlators,” Opt.Eng. 37, P. 52 (1998).
[25]F. T. J. Johnson, T. H. Barnes, T. Eiju, T. G. Haskell and K. Matsuda, “Analysis of a joint transform correlation using a phase-only spatial light modulator,” Opt. Eng. 30, P. 1947 (1991).
[26]Chun-Te Li and S. Yin, “Joint transform correlator in security applications,” IEEE, 97, P. 77 (1997).
[27]Chun-Te Li, S. Yin and F. T. S. Yu, “Nonzero-order joint transform correlator,” Opt. Eng. 37, P. 58 (1998).
[28]R. Tripathi and K. Singh, “Pattern discrimnation using a bank of wavelet filters in a joint transform correlator,” Opt. Eng. 37, P. 532 (1998).
[29]B. Javidi and C. J. Kuo, “Joint transform image correlation using a binary spatial modulator at the Fourier plane,” Appl. Opt. 27, P. 663 (1988).
[30]B. Javidi, J. Wang, and Q. Tang, “Multi-Object binary joint transform correlation using multiple-level thresholding crossing,” Appl. Opt. 30, P. 4234 (1991).
[31]B. Javidi, and J. Wang, “Binary nonlinear joint transform correlation with median and submedian thresholding,” Appl. Opt. 30, P. 967 (1991).
[32]A. Carnicer, I. Juvells and S. Vallmitjana, “Effects of thresholding level variation in fringe binarization of multiobject joint transform correlation,” Appl. Opt. 31, P. 1012 (1992).
[33]B. Javidi, and Q. Tang, “Binary encoding of gray-scale nonlinearly transformed filters for optical pattern recognition,” Appl. Opt. 31, P. 4034 (1992).
[34]R. K. Wang, L. Shang, and C. R. Chatwin, “Cosine wave encoded joint transform correlation,” Opt. Eng. 35, P. 1901 (1996).
[35]B. V. K. V. Kumar and L. Hassebrook, “Performance measure for correlation filters,” Appl. Opt. 29, P. 2997 (1990).
[36]J. L. Horner, “Metrics for assessing pattern-recognition performance,” Appl. Opt. 31, P. 165 (1992).
[37]Yuan-Neng Hsu and H. H. Arsenault, “Optical pattern recognition using circular harmonic expansion,” Appl. Opt. 21, P. 4016 (1982).