基于生物视觉的图像匹配研究
|
摘要
基于生物视觉原理的图像匹配技术,由于其运算速度快、精度高、受制约小,已经受到人们越来越多的重视。新一代的无人机,以及导航设备,很多已经采用了基于生物视觉原理的景象匹配导航技术。
但是对于景象匹配导航而言,仅采用可见光图像匹配制导是不够的,在有云雾遮挡的条件下,或者是天气状况很差的情况,可见光图像匹配制导方法就受到了很大的制约,异源图像匹配制导方法可以有效的解决这个问题。
本文包括基于生物视觉的可见光图像匹配工作,以及异源图像(主要是SAR图像与可见光图像)匹配工作。
本文主要内容有三部分：
1、研究了可见光图像匹配的前期预处理方法,主要包括图像的几何校正以及图像的灰度处理。
2、研究并实现了可见光图像匹配的相关技术,主要包括低矮目标匹配和立体图像匹配,提出了立体效应抹除技术,并且提出了亮饱和图像处理方法。
3、研究了可见光图像与SAR图像的匹配,实现了基于自适应迭代滤波的可见光图像与SAR图像匹配方法和基于梯度的可见光图像与SAR图像匹配方法。
The Image Matching from biology vision attached more and more attentions because of its higher operation speed, higher precision, less restriction. The new unpiloted aircraft and navigation device has adopted the technology of vision navigation. For Image Matching Navigation, it's not enough for visible light image.
In case of the mist keeping out, or the condition of weather is very bad, the method of visible light image matching is more restricted. It's very important to invent a new method of heterogenous image matching. The heterogenous image mainly included infrared image, SAR image.
This paper is for Image Matching from biology vision, which included correlative experiment about the Image Matching, and SAR image and visible light image matching.
The paper is mainly including three parts:
1. The work of preoperation of the visible light image before image matching:It's including The model of geometry correction, the gray-operation model.
2. The research of the visible light image matching. Introducing a lot of basic concept of the visible light image matching, the principle of image matching from biology vision. We have do abundant experiment about matching for lower area, matching for area of severity three-dimensional effect, matching for area of non-severity three-dimensional effect, matching for area of light-saturation. The result of matching is very ideal from abundant experiment.
3. The research of SAR image matching, Achieving A method of matching of the visible image and the SAR image basing on self-adapting iterative filter, and A method of matching of the visible image and the SAR image basing on grads.
但是对于景象匹配导航而言,仅采用可见光图像匹配制导是不够的,在有云雾遮挡的条件下,或者是天气状况很差的情况,可见光图像匹配制导方法就受到了很大的制约,异源图像匹配制导方法可以有效的解决这个问题。
本文包括基于生物视觉的可见光图像匹配工作,以及异源图像(主要是SAR图像与可见光图像)匹配工作。
本文主要内容有三部分：
1、研究了可见光图像匹配的前期预处理方法,主要包括图像的几何校正以及图像的灰度处理。
2、研究并实现了可见光图像匹配的相关技术,主要包括低矮目标匹配和立体图像匹配,提出了立体效应抹除技术,并且提出了亮饱和图像处理方法。
3、研究了可见光图像与SAR图像的匹配,实现了基于自适应迭代滤波的可见光图像与SAR图像匹配方法和基于梯度的可见光图像与SAR图像匹配方法。
The Image Matching from biology vision attached more and more attentions because of its higher operation speed, higher precision, less restriction. The new unpiloted aircraft and navigation device has adopted the technology of vision navigation. For Image Matching Navigation, it's not enough for visible light image.
In case of the mist keeping out, or the condition of weather is very bad, the method of visible light image matching is more restricted. It's very important to invent a new method of heterogenous image matching. The heterogenous image mainly included infrared image, SAR image.
This paper is for Image Matching from biology vision, which included correlative experiment about the Image Matching, and SAR image and visible light image matching.
The paper is mainly including three parts:
1. The work of preoperation of the visible light image before image matching:It's including The model of geometry correction, the gray-operation model.
2. The research of the visible light image matching. Introducing a lot of basic concept of the visible light image matching, the principle of image matching from biology vision. We have do abundant experiment about matching for lower area, matching for area of severity three-dimensional effect, matching for area of non-severity three-dimensional effect, matching for area of light-saturation. The result of matching is very ideal from abundant experiment.
3. The research of SAR image matching, Achieving A method of matching of the visible image and the SAR image basing on self-adapting iterative filter, and A method of matching of the visible image and the SAR image basing on grads.
引文
[1]陈东,李飚,沈振康,SAR与可见光图象匹配方法的研究[J].中国图象图形学报,2001,6(12)：57-59
[2]左峥嵘,杨卫东,张天序,基于空间关系约束的雷达景象匹配算法研究[J].华中科技大学学报(自然科学版),2004,8：168-170
[3]Li Tiejun,Chen Zhe,"Fuzzy feature matching between infrared image and optical image"[J],Proc.SPIE Vol.4602,2001:255-258
[4]Huang Xishan,CHEN Zhe,"A Wavelet — BasedMultisensor Image Registration Algorithm"[J].IEEE.Trans,ICSP'02 Proceedings,2002:144-146
[5]Hui Li,B.S.Manjunath,Sanjit K. Mitra,"A Contour-Based Approach to Multisensor Image Registration"[J],IEEE Trans.Image Processing,1995,4(3):320-334
[6]Jordi INGLADA.Frederic ADRAGNA."AutomaticMultiSensor Image Registration by Edge Matching using Genetic Algorithms"[C].IEEE.Trans.International Geosci2ence and Remote Sensing Symposium,IGARSS,2001:233-234
[7]马俊,曹治国,基于边缘信息的红外与可见光图像匹配技术[J],计算机与数字工程,2006,34(12)：30-32
[8]郭勤,景象匹配技术发展概述[J]．红外与激光工程.2007,36：57-61
[9]Wiley C A. Synthetic aperture radars-a paradigm for technology evolution [J].IEEE Transactions on Aerospace and Electrical System[J],AES-1951,21(3):440-443
[10]Sherwin C W,RuinaJP,Rawcliffe R D.Some early developments in syntheticapert-ure radar systems[J].IEETrans.On Military Elec.,1962,7(1)6:11-115
[11]Cutrona L J,Vivian W E, Leith E,et,al. A high resolution radar combat surveillance system[J]. IRE Trans. Military Elec,1961,6(1):127-131
[12]Kirk JC, Jr. Digital synthetic aperture radar technology[C]. IEEE International Radar Conference Record.1975,2(1):482-487
[13]BennettJR,Widmer P,CummingI G. Areal-time airborne SAR processor[J].Proc.ES-A Working Group, Erascati, Italy,1980,25(2):1024-1031
[14]F.I.Gonzalez,R.C.Beal,etal.Seasat Synthetic Aperture Radar[J]:Ocean Wave Detect-ion Capabilities Science.1979:189-190
[15]星载雷达应用研究[M].国家高技术信息获取与处理技术主题,1996：356-358
[16]张俊,合成孔径雷达(SAR)图像处理与分析方法研究[D],武汉,华中理工大学博士论文,1998：455-457
[17]黄春艳,杨国胜,基于Gabor小波和主元分析的纹理图像分割[J],多媒体技术,2006,32(313),46-49
[18]汪洪桥,赵宗涛,蔡艳宁,方文泯.一种有效增强景象匹配制导图适配性的方法[J].制导与引信,2004,25(4)：10-14
[19]王永明。一种基于小波的快速景象预处理算法[J],宇航学报,2005.3,26(2)：180-184
[20]张维胜,王超,王永明,张红,张波,吴樊,随动系统自动几何精纠正模型与算法实现[J],遥感学报,2005,9(4)：393-397
[21]孙即祥.图像处理[M].北京：科学出版社：150-155
[22]Brown Lisa G.A Survey of Image Registration Techniques[J].ACM Computing Surveys,1992,24(4):325-376
[23]赵锋伟,李吉成,沈振康,景象匹配技术研究,系统工程与电子技术[M],2002,12(24)：111-114
[24]Po M, Cheung C K, Ma W C. A Novel Fast Block Motion Estimation Algorithm Using Center-Biased Search Pattern [C]. Proc. IEEE Int. Conf. on Neural Net and Signal,1995 (2):1616-1619
[25]Mount David M, Netanyahu Nathan S, Moigne Jacqueline Le. Efficient Algorithm for Robust Feature Matching [J].Pattern Recognition,1999,32:17-38
[26]Dai X L,Khorram Siamak. Development of a Feature-Based Approach to Automated Image Registration for Multitemporal and Multisensor Remotely SensedImagery[C]. Proceedings of the 1997 IEEE International Geoscience and Remoth Sensing Sympo-sium,1997:243-245
[27]桑农,张天序.旋转与比例不变的点特征松弛匹配算法[J].电子学报,1998,26(6)： 74-77
[28]IsenorDK.Fingerprint Identification Using GraphMatching[J].Pattern Recogniti-on,1986,19 (2):113-122
[29]刘扬,赵峰伟,金善良,景象匹配区选择方法研究[J],红外与激光工程,2001, 30 (3):168-170
[30]Xiao Yijin,Ding Mingyue,Zhou Chengping.Matching suitability analysis of reference map for scene matching based navigation[A].SPIE[C].1998,3545.02272786X.
[31]Bailey H H,Blackwell F W, Lowery C L,etal.Image correlation,Part Ⅰ:simulation and analysis [R].1976:246-248
[32]沈振康,地图(图像)匹配制导技术[M].长沙：国防科技大学出版社,1983
[33]刘扬,李兵.基于互相关峰特征的景象匹配性能预测方法[J],战术导弹技术,2001,3：54-57
[34]谢亚滨,张万清.景象匹配中重复模式的分析[J].红外与激光工程,1997,26(6).4-7
[35]Ratkovic J A.Estimation techniques and other work on image correlation[R].AD-A046577.1997:236-238
[36]刘扬,赵峰伟.景象匹配区选择方法研究[J].红外与激光工程,2001,30(3)：168-170
[37]Wessely H W.Image correlation:Part Ⅱ:Theoretical basis[R].AD-A036482.1976:6-8
[38]张金槐,蔡洪,飞行器试验统计学[M].国防科技大学出版社：175-177
[39]孙家抦主编,遥感原理与应用[M].武汉大学出版社,2003：354-356
[40]Shi,Z.and Fung.K.B.A Comparison of Digital Speckle Filters[J].Proceedings of IGR-ASS'94.August 1994.Pasadena:8-10
[41]Li,C.Two Adaptive Filters for Speckle Reduction in SAR I magery by Using the Variance Ratio[J].International Journal of Remote Sensing.1987,Vol.9(4).641-653
[42]JI NGFENGXIAO, JINGLI,and A.M OODY.A detail-preserving and flexible ad-aptive filter for speckle suppression in SAR imagery[J].INT.J.Remote sensing.2003.V-OL.24,NO.12.2451-2465
[2]左峥嵘,杨卫东,张天序,基于空间关系约束的雷达景象匹配算法研究[J].华中科技大学学报(自然科学版),2004,8：168-170
[3]Li Tiejun,Chen Zhe,"Fuzzy feature matching between infrared image and optical image"[J],Proc.SPIE Vol.4602,2001:255-258
[4]Huang Xishan,CHEN Zhe,"A Wavelet — BasedMultisensor Image Registration Algorithm"[J].IEEE.Trans,ICSP'02 Proceedings,2002:144-146
[5]Hui Li,B.S.Manjunath,Sanjit K. Mitra,"A Contour-Based Approach to Multisensor Image Registration"[J],IEEE Trans.Image Processing,1995,4(3):320-334
[6]Jordi INGLADA.Frederic ADRAGNA."AutomaticMultiSensor Image Registration by Edge Matching using Genetic Algorithms"[C].IEEE.Trans.International Geosci2ence and Remote Sensing Symposium,IGARSS,2001:233-234
[7]马俊,曹治国,基于边缘信息的红外与可见光图像匹配技术[J],计算机与数字工程,2006,34(12)：30-32
[8]郭勤,景象匹配技术发展概述[J]．红外与激光工程.2007,36：57-61
[9]Wiley C A. Synthetic aperture radars-a paradigm for technology evolution [J].IEEE Transactions on Aerospace and Electrical System[J],AES-1951,21(3):440-443
[10]Sherwin C W,RuinaJP,Rawcliffe R D.Some early developments in syntheticapert-ure radar systems[J].IEETrans.On Military Elec.,1962,7(1)6:11-115
[11]Cutrona L J,Vivian W E, Leith E,et,al. A high resolution radar combat surveillance system[J]. IRE Trans. Military Elec,1961,6(1):127-131
[12]Kirk JC, Jr. Digital synthetic aperture radar technology[C]. IEEE International Radar Conference Record.1975,2(1):482-487
[13]BennettJR,Widmer P,CummingI G. Areal-time airborne SAR processor[J].Proc.ES-A Working Group, Erascati, Italy,1980,25(2):1024-1031
[14]F.I.Gonzalez,R.C.Beal,etal.Seasat Synthetic Aperture Radar[J]:Ocean Wave Detect-ion Capabilities Science.1979:189-190
[15]星载雷达应用研究[M].国家高技术信息获取与处理技术主题,1996：356-358
[16]张俊,合成孔径雷达(SAR)图像处理与分析方法研究[D],武汉,华中理工大学博士论文,1998：455-457
[17]黄春艳,杨国胜,基于Gabor小波和主元分析的纹理图像分割[J],多媒体技术,2006,32(313),46-49
[18]汪洪桥,赵宗涛,蔡艳宁,方文泯.一种有效增强景象匹配制导图适配性的方法[J].制导与引信,2004,25(4)：10-14
[19]王永明。一种基于小波的快速景象预处理算法[J],宇航学报,2005.3,26(2)：180-184
[20]张维胜,王超,王永明,张红,张波,吴樊,随动系统自动几何精纠正模型与算法实现[J],遥感学报,2005,9(4)：393-397
[21]孙即祥.图像处理[M].北京：科学出版社：150-155
[22]Brown Lisa G.A Survey of Image Registration Techniques[J].ACM Computing Surveys,1992,24(4):325-376
[23]赵锋伟,李吉成,沈振康,景象匹配技术研究,系统工程与电子技术[M],2002,12(24)：111-114
[24]Po M, Cheung C K, Ma W C. A Novel Fast Block Motion Estimation Algorithm Using Center-Biased Search Pattern [C]. Proc. IEEE Int. Conf. on Neural Net and Signal,1995 (2):1616-1619
[25]Mount David M, Netanyahu Nathan S, Moigne Jacqueline Le. Efficient Algorithm for Robust Feature Matching [J].Pattern Recognition,1999,32:17-38
[26]Dai X L,Khorram Siamak. Development of a Feature-Based Approach to Automated Image Registration for Multitemporal and Multisensor Remotely SensedImagery[C]. Proceedings of the 1997 IEEE International Geoscience and Remoth Sensing Sympo-sium,1997:243-245
[27]桑农,张天序.旋转与比例不变的点特征松弛匹配算法[J].电子学报,1998,26(6)： 74-77
[28]IsenorDK.Fingerprint Identification Using GraphMatching[J].Pattern Recogniti-on,1986,19 (2):113-122
[29]刘扬,赵峰伟,金善良,景象匹配区选择方法研究[J],红外与激光工程,2001, 30 (3):168-170
[30]Xiao Yijin,Ding Mingyue,Zhou Chengping.Matching suitability analysis of reference map for scene matching based navigation[A].SPIE[C].1998,3545.02272786X.
[31]Bailey H H,Blackwell F W, Lowery C L,etal.Image correlation,Part Ⅰ:simulation and analysis [R].1976:246-248
[32]沈振康,地图(图像)匹配制导技术[M].长沙：国防科技大学出版社,1983
[33]刘扬,李兵.基于互相关峰特征的景象匹配性能预测方法[J],战术导弹技术,2001,3：54-57
[34]谢亚滨,张万清.景象匹配中重复模式的分析[J].红外与激光工程,1997,26(6).4-7
[35]Ratkovic J A.Estimation techniques and other work on image correlation[R].AD-A046577.1997:236-238
[36]刘扬,赵峰伟.景象匹配区选择方法研究[J].红外与激光工程,2001,30(3)：168-170
[37]Wessely H W.Image correlation:Part Ⅱ:Theoretical basis[R].AD-A036482.1976:6-8
[38]张金槐,蔡洪,飞行器试验统计学[M].国防科技大学出版社：175-177
[39]孙家抦主编,遥感原理与应用[M].武汉大学出版社,2003：354-356
[40]Shi,Z.and Fung.K.B.A Comparison of Digital Speckle Filters[J].Proceedings of IGR-ASS'94.August 1994.Pasadena:8-10
[41]Li,C.Two Adaptive Filters for Speckle Reduction in SAR I magery by Using the Variance Ratio[J].International Journal of Remote Sensing.1987,Vol.9(4).641-653
[42]JI NGFENGXIAO, JINGLI,and A.M OODY.A detail-preserving and flexible ad-aptive filter for speckle suppression in SAR imagery[J].INT.J.Remote sensing.2003.V-OL.24,NO.12.2451-2465