基于最大平均相关高度算法的畸变目标识别
|
摘要
应用光电混合相关器对复杂背景下的目标进行探测是目标探测领域的前沿技术。光电混合相关器将光学系统的高速、并行性和电子系统的可控、可编程性相结合,具有识别速度快、定位精度高等优点。光电混合相关器主要包括两种形式,即联合变换相关器和Vander Lugt相关器。目标和模板在联合变换相关器上的识别结果是一对相关峰输出,这对相关峰的相对位置代表目标及模板的相对位置,因此通过确定相关峰的位置,可以确定目标的位置。目标在Vander Lugt相关器上的识别结果是一个相关峰输出,由相关峰的位置可以确定目标位置。
光学相关器是基于相关匹配原理进行目标识别的,只有当目标和预先设定的模板完全一致时才会有相关峰输出。如果目标相对模板产生畸变(比例畸变或旋转畸变),相关峰会迅速减弱导致目标无法识别定位。
本论文从综合鉴别函数入手,深入研究了畸变目标在光学相关器上的识别技术。首先通过优化已有的最大平均相关高度滤波器在Vander Lugt相关器上实现了比例不变识别或旋转不变识别。
而后提出将小波带通滤波器和最大平均相关高度算法相结合,设计了墨西哥帽-最大平均相关高度滤波器,扩大了单个滤波器在Vander Lugt相关器上的比例畸变容差或旋转畸变容差。
然后提出将Sobel边缘检测器和最大平均相关高度算法相结合,改进了最大平均相关高度滤波器的性能,进一步扩大了单个滤波器在Vander Lugt相关器上的比例畸变容差或旋转畸变容差。并且结合Sobel边缘检测器设计了最大平均相关高度参考模板,在联合变换相关器上实现了比例不变识别或旋转不变识别。
最后提出将高斯频域低通滤波器和最大平均相关高度算法相结合,设计了高斯-最大平均相关高度滤波器,使用单个滤波器在Vander Lugt相关器上实现了混合畸变目标的识别(目标相对模板同时存在比例畸变和旋转畸变)。
本论文对所设计的滤波器进行了计算机仿真实验,对所设计的最大平均相关高度参考模板进行了光学识别实验。实验结果表明,优化后的最大平均相关高度滤波器在Vander Lugt相关器上的比例畸变容差为0.76-1.36倍,旋转畸变容差为-15-20度；所设计的墨西哥帽-最大平均相关高度滤波器在Vander Lugt相关器上的比例畸变容差为0.79-1.46倍,旋转畸变容差为-35-35度：通过Sobel边缘检测器改进的最大平均相关高度滤波器在Vander Lugt相关器上的比例畸变容差为0.76-1.56倍,旋转畸变容差范围为80度,而最大平均相关高度参考模板在联合变换相关器上的比例畸变容差为0.93～1.21倍,旋转畸变容差范围为20度；所设计的高斯-最大平均相关高度滤波器在Vander Lugt相关器上能够实现混合畸变目标识别,所得畸变容差为旋转-8～24度,比例缩放0.92～1.16倍。
Detection of targets with cluttered background by hybrid optoelectronics correlator is the front technology in this field. The hybrid optoelectronics correlator combines the high-speed and the parallelism of the optical system with the controllable and programmable of the electronic system, so it has the advantage of high speed recognition and high positioning accuracy. The hybrid optoelectronics correlators include two forms, namely Joint transform correlator (JTC) and Vander Lugt Correlator (VLC). In the JTC, the recognition result of the target and template is a pair of the correlation peak, the location of the pair of correlation peak represents the relative position of the target and template. Therefore, by determining the location of the correlation peak, the position of the target can be determined. In the VLC, the recognition result of the target is a correlation peak. The target position can be determined by the position of the correlation peak.
The target recognized by optical correlator is based on the principle of correlation matching. Only when the detected target is same as the pre-established template, the correlation peak will be output. If the target has the distortion (scale distortion or rotation distortion) relatively the template, the correlation peak will weaken rapidly, so that the target cannot be recognized.
In this paper, started from the synthetic discriminant function, the distortion target recognition technology in the optical correlator is further researched. First, through the optimization of the existing Maximum Average Correlation Height filter, the scale invariant or rotation invariant recognition is achieved in the VLC.
Second, the method of combining the wavelet band-pass filter with the Maximum Average Correlation Height algorithm is proposed, the Mexican Hat Maximum Average Correlation Height filter is designed. The scale distortion or rotation distortion tolerance of a single filter in the VLC is expanded.
Third, the method of combining the Sobel edge detector with the Maximum Average Correlation Height algorithm is proposed, the performance of the Maximum Average Correlation Height filter is improved. The scale distortion or rotation distortion tolerance of a single filter in the VLC is further expanded. Combining the Sobel edge detector, the Maximum Average Correlation Height reference template is designed. The scale invariant or rotation invariant recognition is achieved in JTC using the Maximum Average Correlation Height reference template.
Finally, the method of combining the Gauss frequency domain low-pass filter with the Maximum Average Correlation Height algorithm is proposed, the Gauss Maximum Average Correlation Height filter is designed. The recognition of Hybrid distortion target (existing the scale distortion and rotation distortion simultaneously) in VLC is achieved by a single filter.
In this paper, the designed filters are completed the computer simulation experiment, the designed Maximum Average Correlation Height reference template is completed the optical recognition experiment. The experimental results show that the scale distortion tolerance of the optimized Maximum Average Correlation Height filter in VLC is0.76～1.36times, the rotation distortion tolerance is-15～20degrees. The scale distortion tolerance of the designed Mexican Hat Maximum Average Correlation Height filter in VLC is0.79～1.46times, the rotation distortion tolerance is-35～35degrees. The scale distortion tolerance of the improved Maximum Average Correlation Height filter by Sobel edge detector is0.76～1.56times, the rotation distortion tolerance range is80degrees. The scale distortion tolerance of Maximum Average Correlation Height reference template in JTC is0.93～1.21times, the rotation distortion tolerance range is20degrees. The designed Gauss Maximum Average Correlation Height filter can recognize the hybrid distortion target with the VLC, the distortion tolerance is rotation-8～24degrees and scaling0.92～1.16times.
光学相关器是基于相关匹配原理进行目标识别的,只有当目标和预先设定的模板完全一致时才会有相关峰输出。如果目标相对模板产生畸变(比例畸变或旋转畸变),相关峰会迅速减弱导致目标无法识别定位。
本论文从综合鉴别函数入手,深入研究了畸变目标在光学相关器上的识别技术。首先通过优化已有的最大平均相关高度滤波器在Vander Lugt相关器上实现了比例不变识别或旋转不变识别。
而后提出将小波带通滤波器和最大平均相关高度算法相结合,设计了墨西哥帽-最大平均相关高度滤波器,扩大了单个滤波器在Vander Lugt相关器上的比例畸变容差或旋转畸变容差。
然后提出将Sobel边缘检测器和最大平均相关高度算法相结合,改进了最大平均相关高度滤波器的性能,进一步扩大了单个滤波器在Vander Lugt相关器上的比例畸变容差或旋转畸变容差。并且结合Sobel边缘检测器设计了最大平均相关高度参考模板,在联合变换相关器上实现了比例不变识别或旋转不变识别。
最后提出将高斯频域低通滤波器和最大平均相关高度算法相结合,设计了高斯-最大平均相关高度滤波器,使用单个滤波器在Vander Lugt相关器上实现了混合畸变目标的识别(目标相对模板同时存在比例畸变和旋转畸变)。
本论文对所设计的滤波器进行了计算机仿真实验,对所设计的最大平均相关高度参考模板进行了光学识别实验。实验结果表明,优化后的最大平均相关高度滤波器在Vander Lugt相关器上的比例畸变容差为0.76-1.36倍,旋转畸变容差为-15-20度；所设计的墨西哥帽-最大平均相关高度滤波器在Vander Lugt相关器上的比例畸变容差为0.79-1.46倍,旋转畸变容差为-35-35度：通过Sobel边缘检测器改进的最大平均相关高度滤波器在Vander Lugt相关器上的比例畸变容差为0.76-1.56倍,旋转畸变容差范围为80度,而最大平均相关高度参考模板在联合变换相关器上的比例畸变容差为0.93～1.21倍,旋转畸变容差范围为20度；所设计的高斯-最大平均相关高度滤波器在Vander Lugt相关器上能够实现混合畸变目标识别,所得畸变容差为旋转-8～24度,比例缩放0.92～1.16倍。
Detection of targets with cluttered background by hybrid optoelectronics correlator is the front technology in this field. The hybrid optoelectronics correlator combines the high-speed and the parallelism of the optical system with the controllable and programmable of the electronic system, so it has the advantage of high speed recognition and high positioning accuracy. The hybrid optoelectronics correlators include two forms, namely Joint transform correlator (JTC) and Vander Lugt Correlator (VLC). In the JTC, the recognition result of the target and template is a pair of the correlation peak, the location of the pair of correlation peak represents the relative position of the target and template. Therefore, by determining the location of the correlation peak, the position of the target can be determined. In the VLC, the recognition result of the target is a correlation peak. The target position can be determined by the position of the correlation peak.
The target recognized by optical correlator is based on the principle of correlation matching. Only when the detected target is same as the pre-established template, the correlation peak will be output. If the target has the distortion (scale distortion or rotation distortion) relatively the template, the correlation peak will weaken rapidly, so that the target cannot be recognized.
In this paper, started from the synthetic discriminant function, the distortion target recognition technology in the optical correlator is further researched. First, through the optimization of the existing Maximum Average Correlation Height filter, the scale invariant or rotation invariant recognition is achieved in the VLC.
Second, the method of combining the wavelet band-pass filter with the Maximum Average Correlation Height algorithm is proposed, the Mexican Hat Maximum Average Correlation Height filter is designed. The scale distortion or rotation distortion tolerance of a single filter in the VLC is expanded.
Third, the method of combining the Sobel edge detector with the Maximum Average Correlation Height algorithm is proposed, the performance of the Maximum Average Correlation Height filter is improved. The scale distortion or rotation distortion tolerance of a single filter in the VLC is further expanded. Combining the Sobel edge detector, the Maximum Average Correlation Height reference template is designed. The scale invariant or rotation invariant recognition is achieved in JTC using the Maximum Average Correlation Height reference template.
Finally, the method of combining the Gauss frequency domain low-pass filter with the Maximum Average Correlation Height algorithm is proposed, the Gauss Maximum Average Correlation Height filter is designed. The recognition of Hybrid distortion target (existing the scale distortion and rotation distortion simultaneously) in VLC is achieved by a single filter.
In this paper, the designed filters are completed the computer simulation experiment, the designed Maximum Average Correlation Height reference template is completed the optical recognition experiment. The experimental results show that the scale distortion tolerance of the optimized Maximum Average Correlation Height filter in VLC is0.76～1.36times, the rotation distortion tolerance is-15～20degrees. The scale distortion tolerance of the designed Mexican Hat Maximum Average Correlation Height filter in VLC is0.79～1.46times, the rotation distortion tolerance is-35～35degrees. The scale distortion tolerance of the improved Maximum Average Correlation Height filter by Sobel edge detector is0.76～1.56times, the rotation distortion tolerance range is80degrees. The scale distortion tolerance of Maximum Average Correlation Height reference template in JTC is0.93～1.21times, the rotation distortion tolerance range is20degrees. The designed Gauss Maximum Average Correlation Height filter can recognize the hybrid distortion target with the VLC, the distortion tolerance is rotation-8～24degrees and scaling0.92～1.16times.
引文
[1]陶纯堪,陶纯匡.光学信息论.北京：科学出版社.2004
[2]宋丰华.现代空间光电信息处理技术及应用.北京：国防工业出版社.2004
[3]宋丰华.现代空间光电系统及应用.北京：国防工业出版社.2004
[4]J. W. Goodman. Introduction to Fourier optics. McGran-Hill,1968.
[5]王文生,郭阳血,张晔.光电混合联合变换相关器目标探测与识别.仪器仪表学报.2002,23(5)：50-52.
[6]刘立欣,刘亦菲,张宇.应用光电联合变换相关器实现目标识别.仪器仪表学报,2006,27(6)：2259-2260.
[7]李修建.用于光学模式识别的光学相关器研究.长沙：国防科学技术大学.2002.
[8]Tien-Hsin Chao. Hanying Zhou, George Reyes.512X512 high-speed grayscale optical correlator. Proceedings of SPIE on Optical Pattern Recognition Ⅺ.2000, Vol.4043:40-45.
[9]Kuo C J, Luo Y. Generalized joint fractional Fourier transform correlators:a compact approach. Appl. Opt,1998, 37(35):681-684.
[10]Arun K.. Gupta, Naveen K. Nishchal, Vinod K. Beri. A hybrid digital-optical correlator for automatic target recognition. Proc. Of SPIE on Optical Pattern Recognition ⅩⅧ,2007, Vol.6574:657406-1-17.
[11]邹听,郎琪,王文生.基于联合变换相关器目标探测的红外导引头光学系统研究.测试技术学报,2008,22(6)：31-34.
[12]Tien-Hsin Chao, George Reyes, Hanying Zhou. Automatic target recognition field demonstration using a grayscale optical correlator. Part of the SPIE Conference on Optical Pattern Recognition Ⅹ USA:Florida,1999, Vol.3715: 399-406.
[13]Don A. Gregory, James C. Kirsch, Jeffrey A. Loudin. Optical correlators:Optical computing that really works. Proceedings of SPIE on Advances in Optical Information Processing IV,1990, Vol.1296:2-19.
[14]魏宏刚.实时畸变不变光学模式识别的可靠性研究.成都：四川大学,2002.
[15]张延忻,申金媛,张文伟等.光学相关不变性模式识别的局限性及其出路.南开大学学报(自然科学版).1999,32(3)：69-74.
[16]陈宇.畸变不变复杂背景目标相关探测技术研究.长春：长春理工大学.2009.
[17]李小舜,魏宏刚,王竹毅等.光学模式识别中参考目标畸变参数宽容度研究.激光杂志.2001,22(3)：13-15.
[18]谢敬辉,廖宁放,曹良才.傅里叶光学与现代光学基础.北京：北京理工大学出版社.2007.
[19]谢建平,明海,王沛.近代光学基础.北京：高等教育出版社.2006
[20]Wensheng Wang, Yu Chen, Cuiping Liang, Hua Miao. Hybrid optoelectronic joint transform correlator for the recognition of target in cluttered scenes. Proceedings of SPIE on Information Optics and Photonics Technology. USA: Bellingham, Vol.5642:204-212.
[21]Philip Busarow. Abhijit Mahalanobis. Utilization of an optical correlator for automatic target recognition against ballistic missile defense targets. Proceedings of SPIE on Optical Pattern Recognition Ⅺ,2000, Vol.4043:140-148.
[22]Chau-Jern Cheng, Han-Yen Tu. Implementation of Nonzero-order joint transform correlator using interferometric technique. Optical Review,2002.9(5):193-196.
[23]Wangji Yu, Takumi Minemoto. Performances of an all-optical subtracted joint transform correlator that uses a photorefractive crystal. Optical and Quantum Electronics,2000,32(3):367-382.
[24]Hsuan T. Chang, Ching T. T. Chen. Enhanced optical image verification based on joint transform correlator adopting Fourier hologram. Optical Review,2004,11(3):165-169.
[25]袁祥岩,于大勇,陶纯堪.基于位相编码的畸变不变联合变换相关器的设计.光子学报,2002,31(10)：1238-1243.
[26]Vinod K Beri, Naveen K Nishchal, Arun K Gupta, et al.. Design parameters of hybrid correlator. Indian Journal of Pure & Applied Physics,2009, Vol.46:844-851.
[27]任中京.二维付立叶变换光学系统的空间带宽积与颗粒大小分析.应用激光,1994,14(1)：8-10.
[28]宋菲君, S. Jutamulia近代光学信息处理.北京：北京大学出版社.2004
[29]沈学举,王永仲,姚广涛Vander Lugt相关器中几种空间匹配滤波器算法比较.光学技术,2008,34(2)：314-316.
[30]王仕璠.信息光学理论与应用.北京：北京邮电大学出版社.2004
[31]韩利,刘树田,王骐等.分数相关尺度畸变不变模式识别特性研究.中国激光,2001,28(5)：429-434.
[32]韩利,刘树田,王骐等.分数相关畸变不变目标是识别特性研究.光子学报,2000,29(2)：131-136.
[33]宋小彬,赵建林,袁文成等.联合分数变换相关器的分数相关特性分析.光子学报,2005,34(9)：1367-1371.
[34]李俊,王红霞,何俊发等.一种分数相关尺度畸变不变识别新方法.激光技术,2005,29(2)：180-182.
[35]王金婵,赵永安,王西安.应用联合分数傅里叶变换相关器识别多个物体.光子学报,2002,31(6)：681-684.
[36]D. mendlovic, H. M. Ozaktas. Fractional Fourier transforms and their optical implementation. J. Opt. Soc. Am. A, 1993,10(9):1875-1881.
[37]M. A. Kutay, H. M. Ozaktas, O. Arikan et al. Optimal filtering in fractional Fourier domains. IEEE Trans. Sig. Proc., 1997,45:1129-1143.
[38]R. G. Dorsch, A. W. Lohmann, Y. Bitran et al. Chirp filtering in the fractional Fourier domain. Appl. Opt.,1994, 33(22):7599-7602.
[39]V. Namias. The fractional order Fourier transform and its application to quantum mechanics. J. Inst. Maths.1980, 25(4):241-265.
[40]宋小彬.联合分数傅里叶变换相关器的研究.西安：西北工业大学,2005.
[41]罗世尚.极-梅林变换法在联合变换相关目标识别中的应用研究.长沙：国防科学技术大学,2010.
[42]吴友朋,王红霞,周战荣等.圆谐梅林滤波器在模式识别中的计算机仿真研究.激光技术.2007,31(1)：86-88.
[43]陈宇,苗华,郎琪等.梅林变换在光电混合目标探测技术中的应用.中国激光.2009,36(2)：421-425.
[44]程惠全,聂守平,卞松玲.正交傅里叶-梅林矩变换在光学模式识别中的应用.光电工程,1996.23(4)：16-21.
[45]沈克精.梅林(Mellin)变换及其应用.安徽大学学报(自然科学版),1992,3：15-22.
[46]Samuel P, Kozaitis, Sila Thangwaritorn. Circular optimal trade-off and distance-classifier correlation filters. Society of Photo-Optical Instrumentation Engineers,2000,39(5):1162-1168.
[47]黄晓菁,黄献烈.一种选取圆谐展开中心的新方法.厦门大学学报(自然科学版),1993,32(Sup.2)：135-137.
[48]梅文辉,赵达尊.一种多圆谐光学图像识别方法.北京理工大学学报.1991,11(4)：80-85.
[49]刘大禾.用零阶圆谐滤波器实现不变性图像分割.光学学报,1995,15(6)：744-748.
[50]官玖洪,王肇圻,刘宏利等.应用相移技术的圆谐展开联合变换相关.科学通报,2000,25(9)：996-999.
[51]陈学文,陈祯培.旋转不变振幅位相合成圆谐滤波器.光学学报,1996,16(3)：373-377.
[52]黄晓菩,黄献烈.相似图像的旋转不变识别.厦门大学学报(自然科学版).1993,32(3)：299-303.
[53]杨百愚,冯大毅,袁宝玺.最优圆谐滤波器级次的选取方法.光电子技术与信息,2005,18(6)：53-57.
[54]David Ye, Weston Edens, Thomas T. Lu, et al. Neural network target identification system for false alarm reduction. Proceedings of SPIE on Optical Pattern Recognition XX,2009, Vol.7340:73400K-1-11.
[55]陈泽宇,翁默颖,一种新的神经网络畸变不变模式识别方法.华东师范大学学报(自然科学版),1994.2：47-52.
[56]熊毅之,卞松铃,刘明,陶纯堪.平移、旋转及尺度不变光学神经网络识别系统.红外与激光工程,1997,26(5)：17-21.
[57]李英华.两种解决光学相关识别中畸变不变问题的综合鉴别函数性能比较分析.激光杂志,2007,28(6)：62-63.
[58]深学举,王永仲,姚广涛,胡文刚.扩展等相关峰综合鉴别函数算法研究.光学精密工程,2008,16(1)：156-160.
[59]李英华.两种解决光学相关识别中畸变不变问题的综合鉴别函数性能比较分析.激光杂志,2007,28(6)：62-63.
[60]Shoude Chang, Simon Boothroyd, Paparao Palacharla. Rotation-invariant pattern recognition using a joint transform correlator. Optics Communications,1996,127(2):107-116.
[61]Peter Bone, Rupert Young, Chris Chatwin. Position-, rotation-, scale-, and orientation-invariant multiple object recognition from cluttered scenes. Optical Engineering,2006, Vol.45(7):077203-1-8.
[62]Peter Bone, Rupert Young, Chris Chatwin. Position, rotation, scale, and orientation invariant object tracking from cluttered scenes. Proceedings of SPIE on Optical Pattern Recognition XVII, Vol.6245:624508-1-9.
[63]沈学举,王永仲,李英华,陈亚彬.非约束维纳滤波综合鉴别函数算法研究.光子学报,2006,35(4)：630-634.
[64]D. Casasent, W. Rozzi. Computer generated and phase-only synthetic discriminant function filters. App. Opt.,1986, 25(20):3767-3772.
[65]赵昱,申铉国.四元量化的综合鉴别函数实现旋转不变相关识别.仪器仪表学报,2004,25(4)增刊：473-475.
[66]谢敬辉,王旦福,张浩.应用综合鉴别函数实现畸变不变图像识别.北京理工大学学报,2007,27(10)：915-918
[67]甘厚吉,华文深,王金玉.常规综合鉴别函数的改进.应用光学,2007,28(3)：376-379.
[68]Kumar B. V. Tutorial survey of composite filter designs for optical correlators. Applied optics,1992,31(23): 4773-4801.
[69]Casasent D, Chang Wenthong. Correlation synthetic discriminant function. Applied optics,1986,25(14):2343-2350.
[70]陈前荣,周金鹏.用等相关峰或迭代法合成旋转不变综合鉴别函数.光电子·激光.2003,14(3)：299-302
[71]Abhijit Mahalanobis, B. V. K. Vijaya Kumar, David Casasent. Minimum average correlation energy filters. Applied Optics,1987, Vol.26(17):3633-3640.
[72]B. V. K. Vijaya Kumar. Minimum-variance synthetic discriminant functions. Optical Society of America,1986, Vol. 3(10):1579-1584.
[73]薛蕊,王永仲,华文深.最小方差综合鉴别函数的优化算法.光学技术,2006,32(5)：793-798
[74]A. Mahalanobis, Anjana KhunKhun. Purely real unconstrained correlation filters. Proceedings of SPIE on Optical Pattern Recognition Ⅵ,1995:2490/188-193
[75]A. Mahalanobis, B. V. K. Vijaya Kumar, S Song et al. Unconstrained correlation filters. Appl. Opt.1994,33(17): 3751-3759.
[76]邵珺,华文深,周中亮,甘厚吉.一种抗尺度畸变的光学相关图像识别技术.应用光学,2008,29(4)：503-507
[77]David G L. Distinctive image features from scale-invariant key points. Int. J. Comput. Vis.,2004,60(2):91-110.
[78]Henrik, Sjoberg, Bertrand Noharet. Distortion-invariant filter for non-overlapping noise. Applied Optics,1998,37(29): 6922-6930.
[79]赵昱,申铉国.用于旋转不变模式识别的一种逆滤波算法研究.吉林大学学报(信息科学版),2003,21(13)：216-219.
[80]聂守平,卞松铃,刘峰等.三维目标旋转畸变不变光学相关滤波器组的设计.中国激光,1996,23(11)：1001-1005.
[81]安培建,杨爱粉.大范围畸变不变光电混合图像识别研究.应用光学,1994,15(6)：56-59.
[82]王国田,王汝笠,傅艳红等.畸变不变光电混合图像识别技术研究.红外与毫米波学报,2002,21(5)：361-366.
[83]Frank E. Mcfadden. Optimal separable correlation filters. Proceedings of SPIE on Automatic Target Recognition XII, 2002, Vol.4726:82-92.
[84]刘景琳,高万荣,陶纯堪.归一化算法实现畸变不变图像识别研究.光电工程,2006,33(12)：75-78.
[85]陈燕东,刘景琳,孟志强.新型实时光电混合图像识别系统设计.电子测量与仪器学报,2007,21(3)：103-108.
[86]Abhijit Mahalanobis, B. V. K. Vijaya Kumar. Prediction and Comparison of performance of correlation filters. Proceedings of SPIE, Vol.2754:136-141.
[87]Sharif M. A. Bhuiyan, Mohammad S. Alam, Mohaned Alkanhal. Automatic target recognition and tracking in FLIR imagery using extended maximum average correlation height filter and polynomial distance classifier correlation filter. Proceedings of SPIE on Automatic Target Recognition XV, USA:Bellingham,2005, Vol.5807:1-15.
[88]Akber Gardezi, Philip Birch, Ioannis Kypraios, et al. Implementation of the Maximum Average Correlation Height (MACH) filter in the spatial domain for object recognition from clutter backgrounds. Proc. of SPIE on Automatic Target Recognition ⅩⅩ,2010, Vol.7696:7696IS-1-9.
[89]Peter Bone, Rupert Young, Chris Chatwin. Fully invariant multiple object recognition and tracking using MACH and Kalman filters. Proceedings of SPIE on Optical Pattern Recognition XVIII, Vol.6574:65740A-1-11.
[90]Akber Gardezi, Philip Birch, Ioannis Kypraios, et al.. Implementation of the Maximum Average Correlation Height (MACH) filter in the spatial domain for object recognition from clutter backgrounds. Proceedings of SPIE on Automatic Target Recognition ⅩⅩ,2010, Vol.7696:76961S-1-9.
[91]Alan Van Nevel, Abhijit Mahalanobis. Comparative and study of maximum average correlation height filter variants using ladar imagery. Opt. Eng,2003,42(2):541-550.
[92]B. V. K. Vijaya Kumar, Daniel W. Carlson. Optimal trade-off synthetic discriminant function filters for arbitrary devices. Optical Letters,1994, Vol.19(19):1556-1558.
[93]Harrying Zhou, Tien-hsin Chao. MACH filters synthesizing for detecting targets in cluttered environment for gray-scale optical correlator. Part of the SPIE Conference on Optical Pattern Recognition X.1999.394-398.
[94]Firooz Sadjadi, Abhijit Mahalanobis. Adaptive polarimetric tuning for improved radar target classification using MACH filters. Proceedings of SPIE on Automatic Target Recognition XV, USA:Bellingham,2005, Vol.5807:88-99.
[95]Oliver C. Johnson, Weston Edens, Thomas T. Lu, et al.. Optimization of OT-MACH filter generation for target recognition. Proceedings of SPIE on Optical Pattern Recognition ⅩⅩ,2009, Vol.7340:734008-1-9.
[96]Abhijit Mahalanobis, B. V. K. Vijaya Kumar. Optimality of the maximum average correlation height filter for detection of targets in noise. Opt. Eng.1997,36(19):2642-2648.
[97]Jorgen Karlholm. Generalizations of the maximum average correlation height filter. Opt. Soc. Am. A,2000,17(8): 1399-1406.
[98]Abhijt Mahalanobis, B.V.K. Vijaya Kumar, Robert T. Frankot. Intraclass and between-class training-image registration for correlation-filter synthese. Applied Optics,2000,39(17):2918-2924.
[99]陈方涵.小波理论在光学相关探测中的应用研究.长春：长春理工大学,2011.
[100]飞思科技产品研发中心.小波分析理论与MATLAB 7实现.北京：电子工业出版社.2005
[101]谷德军,王东晓,纪忠萍,郑彬.墨西哥帽小波变换的影响域和计算方案新探讨.应用气象学报,2009,20(1)：62-69.
[102]丁南南,刘艳滢,朱明.尺度相互作用墨西哥帽小波提取图像特征点.液晶与显示,2012,27(1)：125-129.
[103]陈跃妤.边缘检测算法比较分析.农业网络信息.2012,6：31-33.
[104]唐路路,张启灿,胡松.一种自适应阈值的Canny边缘检测算法.光电工程,2011,28(5)：127-132.
[105]Harris C, Stephens M. A combined corner and edge detector. Proceedings of the 4th Alvey Vision Conference, UK: Manchester,1988:147-151.
[106]W. Nicholas Greene, Yuhan Zhang, Thomas T. Lu, et al.. Feature extraction and selection strategies for automated target recognition. Proceedings of SPIE on Independent Component Analyses,2010, Vol.7703:77030B-1-11.
[107]冈萨雷斯著,阮秋琦,阮宇智译.数字图像处理.北京：电子工业出版社.2007
[108]胡建建,曾培峰,唐莉萍,臧珠萍.基于高斯低通滤波的音乐节拍提取.东华大学学报,2011,37(1)：72-75.
[2]宋丰华.现代空间光电信息处理技术及应用.北京：国防工业出版社.2004
[3]宋丰华.现代空间光电系统及应用.北京：国防工业出版社.2004
[4]J. W. Goodman. Introduction to Fourier optics. McGran-Hill,1968.
[5]王文生,郭阳血,张晔.光电混合联合变换相关器目标探测与识别.仪器仪表学报.2002,23(5)：50-52.
[6]刘立欣,刘亦菲,张宇.应用光电联合变换相关器实现目标识别.仪器仪表学报,2006,27(6)：2259-2260.
[7]李修建.用于光学模式识别的光学相关器研究.长沙：国防科学技术大学.2002.
[8]Tien-Hsin Chao. Hanying Zhou, George Reyes.512X512 high-speed grayscale optical correlator. Proceedings of SPIE on Optical Pattern Recognition Ⅺ.2000, Vol.4043:40-45.
[9]Kuo C J, Luo Y. Generalized joint fractional Fourier transform correlators:a compact approach. Appl. Opt,1998, 37(35):681-684.
[10]Arun K.. Gupta, Naveen K. Nishchal, Vinod K. Beri. A hybrid digital-optical correlator for automatic target recognition. Proc. Of SPIE on Optical Pattern Recognition ⅩⅧ,2007, Vol.6574:657406-1-17.
[11]邹听,郎琪,王文生.基于联合变换相关器目标探测的红外导引头光学系统研究.测试技术学报,2008,22(6)：31-34.
[12]Tien-Hsin Chao, George Reyes, Hanying Zhou. Automatic target recognition field demonstration using a grayscale optical correlator. Part of the SPIE Conference on Optical Pattern Recognition Ⅹ USA:Florida,1999, Vol.3715: 399-406.
[13]Don A. Gregory, James C. Kirsch, Jeffrey A. Loudin. Optical correlators:Optical computing that really works. Proceedings of SPIE on Advances in Optical Information Processing IV,1990, Vol.1296:2-19.
[14]魏宏刚.实时畸变不变光学模式识别的可靠性研究.成都：四川大学,2002.
[15]张延忻,申金媛,张文伟等.光学相关不变性模式识别的局限性及其出路.南开大学学报(自然科学版).1999,32(3)：69-74.
[16]陈宇.畸变不变复杂背景目标相关探测技术研究.长春：长春理工大学.2009.
[17]李小舜,魏宏刚,王竹毅等.光学模式识别中参考目标畸变参数宽容度研究.激光杂志.2001,22(3)：13-15.
[18]谢敬辉,廖宁放,曹良才.傅里叶光学与现代光学基础.北京：北京理工大学出版社.2007.
[19]谢建平,明海,王沛.近代光学基础.北京：高等教育出版社.2006
[20]Wensheng Wang, Yu Chen, Cuiping Liang, Hua Miao. Hybrid optoelectronic joint transform correlator for the recognition of target in cluttered scenes. Proceedings of SPIE on Information Optics and Photonics Technology. USA: Bellingham, Vol.5642:204-212.
[21]Philip Busarow. Abhijit Mahalanobis. Utilization of an optical correlator for automatic target recognition against ballistic missile defense targets. Proceedings of SPIE on Optical Pattern Recognition Ⅺ,2000, Vol.4043:140-148.
[22]Chau-Jern Cheng, Han-Yen Tu. Implementation of Nonzero-order joint transform correlator using interferometric technique. Optical Review,2002.9(5):193-196.
[23]Wangji Yu, Takumi Minemoto. Performances of an all-optical subtracted joint transform correlator that uses a photorefractive crystal. Optical and Quantum Electronics,2000,32(3):367-382.
[24]Hsuan T. Chang, Ching T. T. Chen. Enhanced optical image verification based on joint transform correlator adopting Fourier hologram. Optical Review,2004,11(3):165-169.
[25]袁祥岩,于大勇,陶纯堪.基于位相编码的畸变不变联合变换相关器的设计.光子学报,2002,31(10)：1238-1243.
[26]Vinod K Beri, Naveen K Nishchal, Arun K Gupta, et al.. Design parameters of hybrid correlator. Indian Journal of Pure & Applied Physics,2009, Vol.46:844-851.
[27]任中京.二维付立叶变换光学系统的空间带宽积与颗粒大小分析.应用激光,1994,14(1)：8-10.
[28]宋菲君, S. Jutamulia近代光学信息处理.北京：北京大学出版社.2004
[29]沈学举,王永仲,姚广涛Vander Lugt相关器中几种空间匹配滤波器算法比较.光学技术,2008,34(2)：314-316.
[30]王仕璠.信息光学理论与应用.北京：北京邮电大学出版社.2004
[31]韩利,刘树田,王骐等.分数相关尺度畸变不变模式识别特性研究.中国激光,2001,28(5)：429-434.
[32]韩利,刘树田,王骐等.分数相关畸变不变目标是识别特性研究.光子学报,2000,29(2)：131-136.
[33]宋小彬,赵建林,袁文成等.联合分数变换相关器的分数相关特性分析.光子学报,2005,34(9)：1367-1371.
[34]李俊,王红霞,何俊发等.一种分数相关尺度畸变不变识别新方法.激光技术,2005,29(2)：180-182.
[35]王金婵,赵永安,王西安.应用联合分数傅里叶变换相关器识别多个物体.光子学报,2002,31(6)：681-684.
[36]D. mendlovic, H. M. Ozaktas. Fractional Fourier transforms and their optical implementation. J. Opt. Soc. Am. A, 1993,10(9):1875-1881.
[37]M. A. Kutay, H. M. Ozaktas, O. Arikan et al. Optimal filtering in fractional Fourier domains. IEEE Trans. Sig. Proc., 1997,45:1129-1143.
[38]R. G. Dorsch, A. W. Lohmann, Y. Bitran et al. Chirp filtering in the fractional Fourier domain. Appl. Opt.,1994, 33(22):7599-7602.
[39]V. Namias. The fractional order Fourier transform and its application to quantum mechanics. J. Inst. Maths.1980, 25(4):241-265.
[40]宋小彬.联合分数傅里叶变换相关器的研究.西安：西北工业大学,2005.
[41]罗世尚.极-梅林变换法在联合变换相关目标识别中的应用研究.长沙：国防科学技术大学,2010.
[42]吴友朋,王红霞,周战荣等.圆谐梅林滤波器在模式识别中的计算机仿真研究.激光技术.2007,31(1)：86-88.
[43]陈宇,苗华,郎琪等.梅林变换在光电混合目标探测技术中的应用.中国激光.2009,36(2)：421-425.
[44]程惠全,聂守平,卞松玲.正交傅里叶-梅林矩变换在光学模式识别中的应用.光电工程,1996.23(4)：16-21.
[45]沈克精.梅林(Mellin)变换及其应用.安徽大学学报(自然科学版),1992,3：15-22.
[46]Samuel P, Kozaitis, Sila Thangwaritorn. Circular optimal trade-off and distance-classifier correlation filters. Society of Photo-Optical Instrumentation Engineers,2000,39(5):1162-1168.
[47]黄晓菁,黄献烈.一种选取圆谐展开中心的新方法.厦门大学学报(自然科学版),1993,32(Sup.2)：135-137.
[48]梅文辉,赵达尊.一种多圆谐光学图像识别方法.北京理工大学学报.1991,11(4)：80-85.
[49]刘大禾.用零阶圆谐滤波器实现不变性图像分割.光学学报,1995,15(6)：744-748.
[50]官玖洪,王肇圻,刘宏利等.应用相移技术的圆谐展开联合变换相关.科学通报,2000,25(9)：996-999.
[51]陈学文,陈祯培.旋转不变振幅位相合成圆谐滤波器.光学学报,1996,16(3)：373-377.
[52]黄晓菩,黄献烈.相似图像的旋转不变识别.厦门大学学报(自然科学版).1993,32(3)：299-303.
[53]杨百愚,冯大毅,袁宝玺.最优圆谐滤波器级次的选取方法.光电子技术与信息,2005,18(6)：53-57.
[54]David Ye, Weston Edens, Thomas T. Lu, et al. Neural network target identification system for false alarm reduction. Proceedings of SPIE on Optical Pattern Recognition XX,2009, Vol.7340:73400K-1-11.
[55]陈泽宇,翁默颖,一种新的神经网络畸变不变模式识别方法.华东师范大学学报(自然科学版),1994.2：47-52.
[56]熊毅之,卞松铃,刘明,陶纯堪.平移、旋转及尺度不变光学神经网络识别系统.红外与激光工程,1997,26(5)：17-21.
[57]李英华.两种解决光学相关识别中畸变不变问题的综合鉴别函数性能比较分析.激光杂志,2007,28(6)：62-63.
[58]深学举,王永仲,姚广涛,胡文刚.扩展等相关峰综合鉴别函数算法研究.光学精密工程,2008,16(1)：156-160.
[59]李英华.两种解决光学相关识别中畸变不变问题的综合鉴别函数性能比较分析.激光杂志,2007,28(6)：62-63.
[60]Shoude Chang, Simon Boothroyd, Paparao Palacharla. Rotation-invariant pattern recognition using a joint transform correlator. Optics Communications,1996,127(2):107-116.
[61]Peter Bone, Rupert Young, Chris Chatwin. Position-, rotation-, scale-, and orientation-invariant multiple object recognition from cluttered scenes. Optical Engineering,2006, Vol.45(7):077203-1-8.
[62]Peter Bone, Rupert Young, Chris Chatwin. Position, rotation, scale, and orientation invariant object tracking from cluttered scenes. Proceedings of SPIE on Optical Pattern Recognition XVII, Vol.6245:624508-1-9.
[63]沈学举,王永仲,李英华,陈亚彬.非约束维纳滤波综合鉴别函数算法研究.光子学报,2006,35(4)：630-634.
[64]D. Casasent, W. Rozzi. Computer generated and phase-only synthetic discriminant function filters. App. Opt.,1986, 25(20):3767-3772.
[65]赵昱,申铉国.四元量化的综合鉴别函数实现旋转不变相关识别.仪器仪表学报,2004,25(4)增刊：473-475.
[66]谢敬辉,王旦福,张浩.应用综合鉴别函数实现畸变不变图像识别.北京理工大学学报,2007,27(10)：915-918
[67]甘厚吉,华文深,王金玉.常规综合鉴别函数的改进.应用光学,2007,28(3)：376-379.
[68]Kumar B. V. Tutorial survey of composite filter designs for optical correlators. Applied optics,1992,31(23): 4773-4801.
[69]Casasent D, Chang Wenthong. Correlation synthetic discriminant function. Applied optics,1986,25(14):2343-2350.
[70]陈前荣,周金鹏.用等相关峰或迭代法合成旋转不变综合鉴别函数.光电子·激光.2003,14(3)：299-302
[71]Abhijit Mahalanobis, B. V. K. Vijaya Kumar, David Casasent. Minimum average correlation energy filters. Applied Optics,1987, Vol.26(17):3633-3640.
[72]B. V. K. Vijaya Kumar. Minimum-variance synthetic discriminant functions. Optical Society of America,1986, Vol. 3(10):1579-1584.
[73]薛蕊,王永仲,华文深.最小方差综合鉴别函数的优化算法.光学技术,2006,32(5)：793-798
[74]A. Mahalanobis, Anjana KhunKhun. Purely real unconstrained correlation filters. Proceedings of SPIE on Optical Pattern Recognition Ⅵ,1995:2490/188-193
[75]A. Mahalanobis, B. V. K. Vijaya Kumar, S Song et al. Unconstrained correlation filters. Appl. Opt.1994,33(17): 3751-3759.
[76]邵珺,华文深,周中亮,甘厚吉.一种抗尺度畸变的光学相关图像识别技术.应用光学,2008,29(4)：503-507
[77]David G L. Distinctive image features from scale-invariant key points. Int. J. Comput. Vis.,2004,60(2):91-110.
[78]Henrik, Sjoberg, Bertrand Noharet. Distortion-invariant filter for non-overlapping noise. Applied Optics,1998,37(29): 6922-6930.
[79]赵昱,申铉国.用于旋转不变模式识别的一种逆滤波算法研究.吉林大学学报(信息科学版),2003,21(13)：216-219.
[80]聂守平,卞松铃,刘峰等.三维目标旋转畸变不变光学相关滤波器组的设计.中国激光,1996,23(11)：1001-1005.
[81]安培建,杨爱粉.大范围畸变不变光电混合图像识别研究.应用光学,1994,15(6)：56-59.
[82]王国田,王汝笠,傅艳红等.畸变不变光电混合图像识别技术研究.红外与毫米波学报,2002,21(5)：361-366.
[83]Frank E. Mcfadden. Optimal separable correlation filters. Proceedings of SPIE on Automatic Target Recognition XII, 2002, Vol.4726:82-92.
[84]刘景琳,高万荣,陶纯堪.归一化算法实现畸变不变图像识别研究.光电工程,2006,33(12)：75-78.
[85]陈燕东,刘景琳,孟志强.新型实时光电混合图像识别系统设计.电子测量与仪器学报,2007,21(3)：103-108.
[86]Abhijit Mahalanobis, B. V. K. Vijaya Kumar. Prediction and Comparison of performance of correlation filters. Proceedings of SPIE, Vol.2754:136-141.
[87]Sharif M. A. Bhuiyan, Mohammad S. Alam, Mohaned Alkanhal. Automatic target recognition and tracking in FLIR imagery using extended maximum average correlation height filter and polynomial distance classifier correlation filter. Proceedings of SPIE on Automatic Target Recognition XV, USA:Bellingham,2005, Vol.5807:1-15.
[88]Akber Gardezi, Philip Birch, Ioannis Kypraios, et al. Implementation of the Maximum Average Correlation Height (MACH) filter in the spatial domain for object recognition from clutter backgrounds. Proc. of SPIE on Automatic Target Recognition ⅩⅩ,2010, Vol.7696:7696IS-1-9.
[89]Peter Bone, Rupert Young, Chris Chatwin. Fully invariant multiple object recognition and tracking using MACH and Kalman filters. Proceedings of SPIE on Optical Pattern Recognition XVIII, Vol.6574:65740A-1-11.
[90]Akber Gardezi, Philip Birch, Ioannis Kypraios, et al.. Implementation of the Maximum Average Correlation Height (MACH) filter in the spatial domain for object recognition from clutter backgrounds. Proceedings of SPIE on Automatic Target Recognition ⅩⅩ,2010, Vol.7696:76961S-1-9.
[91]Alan Van Nevel, Abhijit Mahalanobis. Comparative and study of maximum average correlation height filter variants using ladar imagery. Opt. Eng,2003,42(2):541-550.
[92]B. V. K. Vijaya Kumar, Daniel W. Carlson. Optimal trade-off synthetic discriminant function filters for arbitrary devices. Optical Letters,1994, Vol.19(19):1556-1558.
[93]Harrying Zhou, Tien-hsin Chao. MACH filters synthesizing for detecting targets in cluttered environment for gray-scale optical correlator. Part of the SPIE Conference on Optical Pattern Recognition X.1999.394-398.
[94]Firooz Sadjadi, Abhijit Mahalanobis. Adaptive polarimetric tuning for improved radar target classification using MACH filters. Proceedings of SPIE on Automatic Target Recognition XV, USA:Bellingham,2005, Vol.5807:88-99.
[95]Oliver C. Johnson, Weston Edens, Thomas T. Lu, et al.. Optimization of OT-MACH filter generation for target recognition. Proceedings of SPIE on Optical Pattern Recognition ⅩⅩ,2009, Vol.7340:734008-1-9.
[96]Abhijit Mahalanobis, B. V. K. Vijaya Kumar. Optimality of the maximum average correlation height filter for detection of targets in noise. Opt. Eng.1997,36(19):2642-2648.
[97]Jorgen Karlholm. Generalizations of the maximum average correlation height filter. Opt. Soc. Am. A,2000,17(8): 1399-1406.
[98]Abhijt Mahalanobis, B.V.K. Vijaya Kumar, Robert T. Frankot. Intraclass and between-class training-image registration for correlation-filter synthese. Applied Optics,2000,39(17):2918-2924.
[99]陈方涵.小波理论在光学相关探测中的应用研究.长春：长春理工大学,2011.
[100]飞思科技产品研发中心.小波分析理论与MATLAB 7实现.北京：电子工业出版社.2005
[101]谷德军,王东晓,纪忠萍,郑彬.墨西哥帽小波变换的影响域和计算方案新探讨.应用气象学报,2009,20(1)：62-69.
[102]丁南南,刘艳滢,朱明.尺度相互作用墨西哥帽小波提取图像特征点.液晶与显示,2012,27(1)：125-129.
[103]陈跃妤.边缘检测算法比较分析.农业网络信息.2012,6：31-33.
[104]唐路路,张启灿,胡松.一种自适应阈值的Canny边缘检测算法.光电工程,2011,28(5)：127-132.
[105]Harris C, Stephens M. A combined corner and edge detector. Proceedings of the 4th Alvey Vision Conference, UK: Manchester,1988:147-151.
[106]W. Nicholas Greene, Yuhan Zhang, Thomas T. Lu, et al.. Feature extraction and selection strategies for automated target recognition. Proceedings of SPIE on Independent Component Analyses,2010, Vol.7703:77030B-1-11.
[107]冈萨雷斯著,阮秋琦,阮宇智译.数字图像处理.北京：电子工业出版社.2007
[108]胡建建,曾培峰,唐莉萍,臧珠萍.基于高斯低通滤波的音乐节拍提取.东华大学学报,2011,37(1)：72-75.
© 2004-2018 中国地质图书馆版权所有 京ICP备05064691号 京公网安备11010802017129号 地址：北京市海淀区学院路29号 邮编：100083 电话：办公室：(+86 10)66554848；文献借阅、咨询服务、科技查新：66554700 |