特定目标自动识别方法研究
|
摘要
特定目标的识别技术是模式识别领域非常活跃的研究课题，在矿产资源的探测，气象云图分析天气预报，复杂背景下军事目标的自动识别等方面都有着广泛的应用。本文主要研究了对于特定目标自动识别的方法。
本文首先研究并比较几种基于灰度的匹配算法，即归一化灰度组合相关法，基于局部熵的匹配算法和基于熵的矩的匹配算法，并对其中存在的问题进行了探讨。
然后介绍了基于目标参照物的识别方法，在对停机坪的检测过程中，以跑道为参照物，首先定位机场跑道，然后根据机场跑道与停机坪的位置关系以及停机坪特有的特征来定位停机坪。
最后提出了基于统计分类的目标检测方法，并实现了基于主成分分析的目标识别系统，实验取得了比较理想的结果。
The technology of target recognition is an active subject in the field of pattern recognition. It is widely applied in the field of exploring mineral deposits, analyzing climograph for weather forecast, auto recognition of military target in the complicated background. This paper primarily studies the method of automatic target recognition.
In this paper firstly some gray based matching algorithms, such as normalized correlation algorithm, local entropy based algorithm and moments of entropy based algorithm, are probed. And problems existed in these algorithms are studied.
Then introduces reference object based recognition method. For packing apron detection, we take the airfield runway as the reference object and locate it, then locate the packing apron according to ubiety between the airfield runway and the packing apron and the characteristic of packing apron.
Finally we propose statistical classification based object detecting method, and realize the PCA based object detecting system. The system is experimented and achieves satisfactory results.
本文首先研究并比较几种基于灰度的匹配算法，即归一化灰度组合相关法，基于局部熵的匹配算法和基于熵的矩的匹配算法，并对其中存在的问题进行了探讨。
然后介绍了基于目标参照物的识别方法，在对停机坪的检测过程中，以跑道为参照物，首先定位机场跑道，然后根据机场跑道与停机坪的位置关系以及停机坪特有的特征来定位停机坪。
最后提出了基于统计分类的目标检测方法，并实现了基于主成分分析的目标识别系统，实验取得了比较理想的结果。
The technology of target recognition is an active subject in the field of pattern recognition. It is widely applied in the field of exploring mineral deposits, analyzing climograph for weather forecast, auto recognition of military target in the complicated background. This paper primarily studies the method of automatic target recognition.
In this paper firstly some gray based matching algorithms, such as normalized correlation algorithm, local entropy based algorithm and moments of entropy based algorithm, are probed. And problems existed in these algorithms are studied.
Then introduces reference object based recognition method. For packing apron detection, we take the airfield runway as the reference object and locate it, then locate the packing apron according to ubiety between the airfield runway and the packing apron and the characteristic of packing apron.
Finally we propose statistical classification based object detecting method, and realize the PCA based object detecting system. The system is experimented and achieves satisfactory results.
引文
1．荆仁杰等．计算机图像处理．浙江：浙江大学出版社．1992．
2．赵荣椿．数字图像处理导论．西安：西北工业大学出版社．1995．
3．夏良正．数字图像处理．南京：东南大学出版社．1999．
4．王润生编著．图像理解．国防科技大学出版社．1995．
5．Kenneth. R. Castlenman著．朱志刚等译．数字图像处理．北京：电子工业出版社．
6．吴健康．数字图像分析．北京：人民邮电出版社．1999．
7．李介谷．计算机视觉的理论和实践．上海交通大学出版社．1999．
8．Milan. Sonka、Vaclav. Hlavac、Rager. Boyle．图像处理．分析与机器视觉．人民邮电出版社．2003．
9．蒋明、张桂林．基于主成分分析的图像匹配方法研究．红外与激光工程． 2000．8．Vol．29 No．4．
10．石争浩、黄士坦．遗传算法和BP算法相结合进行图像匹配．武汉大学学报．2003．6．
11．石争浩、黄士坦．遗传算法和BP算法相结合进行图像匹配．武汉大学学报．2003．6．
12．伺仁芳、王乘、杨文兵．基于混沌遗传算法的图像匹配．红外与激光工程．2003．2．
13．严柏俊、郑链、王克勇．基于不变矩特征匹配的快速目标检测算法．红外技术．2002．23(6)．
14．朱音、谢敏．基于目标规格化与zernike矩相结合的目标识别算法．微电子与计算机．2003．1．
15．陈晓钟、孙华燕．基于能量特征的图像目标检测．红外与激光工程．2001．2．
16．高世孩、戴文刚、田仲．空域和频域相结合的景象匹配算法．中国图像图形学报．2000．4．
17. M. Kass、A. Witein、D. Teaopoulos. Active contour models. Int' 1 Journal of Comp Vis. 2002. 1:321-331.
18. K.F. Lai、RT. chin. Deformatble contours:Modeling and extraction. IEEE Trans on Pattern Anal Match Intell. 1995. 17(11):1084-1090.
19．余莉、王润生．基于多尺度变形模板的目标检测与识别．计算机研究与发展．2000．39(10)．
20．钱晓峰、吕伟．基于形状特征划分的树状区域匹配图像检索．南京理工大学学报．2000．25(1)．
21. Brown L.G. A survey of image registration techniques[J]. ACM Computing Surveys 1992. 24(4): 325-376.
22．曹治国、张千．基于局部递归分割的远距红外桥梁识别算法．精确制导技术文集．2003．
23．苏康、关世义．一种实用的归一化互相关景象匹配算法．宇航学报．1997．18(3)：403—407．
24. KUGIinCD、Eppler WG. Map-matching techniques for use with multi Spectral/multitemporal. SPIE Image Processing for Missile Guidance 1999. 238:146-156
25. ShiozaKi A、Edge extraction using entropy operator. CVGIP Pro 1986. 36:1-9.
26. Pal N. et al. Object-background segmentation using new definitions of entropy. IEEE Proc Pt, E. 1999. 136(4):284-295
27．田金文、苏康、柳健．基于局部熵差的图像匹配方法．宇航学报．1999．20(11)．
28. Wang BT、SunJingao. Relative moment and their applications to geometric in shape recognition. Journal of image and graphics. 2001. 6(3):296-300.
29. Hu.M.K. Visual pattern recognition by moment invariants. IRE TruInformation Theory. 1992. IT(8).
30. J.W. Bulter、M.K. Bulter and A. Stroud. "Automatic classification of chromosomes" in K. Enslein, ed. Data Acquisition and Processing in Biology and Medicine 3. Pergamon Press. New York. 1994.
31. Wei Dong min. et al. Using WEB grammer to recognition dimensions in engineering drawings. Pattern Recognition. 1993. 26(9).1407-1416.
2．赵荣椿．数字图像处理导论．西安：西北工业大学出版社．1995．
3．夏良正．数字图像处理．南京：东南大学出版社．1999．
4．王润生编著．图像理解．国防科技大学出版社．1995．
5．Kenneth. R. Castlenman著．朱志刚等译．数字图像处理．北京：电子工业出版社．
6．吴健康．数字图像分析．北京：人民邮电出版社．1999．
7．李介谷．计算机视觉的理论和实践．上海交通大学出版社．1999．
8．Milan. Sonka、Vaclav. Hlavac、Rager. Boyle．图像处理．分析与机器视觉．人民邮电出版社．2003．
9．蒋明、张桂林．基于主成分分析的图像匹配方法研究．红外与激光工程． 2000．8．Vol．29 No．4．
10．石争浩、黄士坦．遗传算法和BP算法相结合进行图像匹配．武汉大学学报．2003．6．
11．石争浩、黄士坦．遗传算法和BP算法相结合进行图像匹配．武汉大学学报．2003．6．
12．伺仁芳、王乘、杨文兵．基于混沌遗传算法的图像匹配．红外与激光工程．2003．2．
13．严柏俊、郑链、王克勇．基于不变矩特征匹配的快速目标检测算法．红外技术．2002．23(6)．
14．朱音、谢敏．基于目标规格化与zernike矩相结合的目标识别算法．微电子与计算机．2003．1．
15．陈晓钟、孙华燕．基于能量特征的图像目标检测．红外与激光工程．2001．2．
16．高世孩、戴文刚、田仲．空域和频域相结合的景象匹配算法．中国图像图形学报．2000．4．
17. M. Kass、A. Witein、D. Teaopoulos. Active contour models. Int' 1 Journal of Comp Vis. 2002. 1:321-331.
18. K.F. Lai、RT. chin. Deformatble contours:Modeling and extraction. IEEE Trans on Pattern Anal Match Intell. 1995. 17(11):1084-1090.
19．余莉、王润生．基于多尺度变形模板的目标检测与识别．计算机研究与发展．2000．39(10)．
20．钱晓峰、吕伟．基于形状特征划分的树状区域匹配图像检索．南京理工大学学报．2000．25(1)．
21. Brown L.G. A survey of image registration techniques[J]. ACM Computing Surveys 1992. 24(4): 325-376.
22．曹治国、张千．基于局部递归分割的远距红外桥梁识别算法．精确制导技术文集．2003．
23．苏康、关世义．一种实用的归一化互相关景象匹配算法．宇航学报．1997．18(3)：403—407．
24. KUGIinCD、Eppler WG. Map-matching techniques for use with multi Spectral/multitemporal. SPIE Image Processing for Missile Guidance 1999. 238:146-156
25. ShiozaKi A、Edge extraction using entropy operator. CVGIP Pro 1986. 36:1-9.
26. Pal N. et al. Object-background segmentation using new definitions of entropy. IEEE Proc Pt, E. 1999. 136(4):284-295
27．田金文、苏康、柳健．基于局部熵差的图像匹配方法．宇航学报．1999．20(11)．
28. Wang BT、SunJingao. Relative moment and their applications to geometric in shape recognition. Journal of image and graphics. 2001. 6(3):296-300.
29. Hu.M.K. Visual pattern recognition by moment invariants. IRE TruInformation Theory. 1992. IT(8).
30. J.W. Bulter、M.K. Bulter and A. Stroud. "Automatic classification of chromosomes" in K. Enslein, ed. Data Acquisition and Processing in Biology and Medicine 3. Pergamon Press. New York. 1994.
31. Wei Dong min. et al. Using WEB grammer to recognition dimensions in engineering drawings. Pattern Recognition. 1993. 26(9).1407-1416.