红外图像的目标检测、识别与跟踪技术研究
|
摘要
随着精确制导武器的发展，对精确制导信息处理技术的要求越来越高，它已成为制约武器发展的关键因素之一。由于红外成像制导具有诸多优点，已成为当今世界精确制导技术发展和研究的主流和方向。本文根据红外成像制导的基本特征、制导信息处理方法研究的发展现状、存在的主要问题以及工程实际应用的基本要求，在红外图像预处理、图像分割、目标识别和跟踪等方面，进行了系统地研究。本文的主要内容如下：
(1)结合热成像技术原理对红外图像的特征进行了分析，重点研究了红外图像的灰度直方图，提出了隔行／列直方图统计的思想。以灰度直方图作为统计手段和依据，提出了自适应灰度分段线性变换的图像增强算法。
(2)研究了一种用于红外成像制导信息处理的自适应、多尺度、抗噪声小波边缘检测方法。该方法利用小波变换的自动变焦的多分辨分析思想，构造了适合的噪声中边缘检测的多分辨率高斯梯度算子。实验结果表明，该方法是一种快速、有效的自动化目标检测方法。
(3)在模式识别理论的基础上，探讨了图像识别系统，研究了不变量(不变矩、圆度)在目标识别中的应用，并针对圆度法用于直升机识别给出了应用实例，提出拓扑思想在识别中应用的方案，并针对坦克提出了一种命中点选择方案和算法。
(4)介绍了成像跟踪常用的两大方法，波门法、匹配法，并在波门法中提出了基于命中点选择的命中点跟踪法，研究了跟踪预测算法的应用，针对红外图像目标提出了一种卡尔曼滤波改进算法。
With the development of precision guided weapons, the requirements on information processing techniques for precision guidance is becoming higher and higher. This has become a key factor, which restricts the development of the weapon. Since IR imaging guidance possesses many advantages, it has become the trend and direction for current development and study of precision guidance technologies worldwide. On the basis of the fundamental characteristics of IR imaging guidance, the current development status of guidance information processing methods, the main problems existed as well as the basic requirements on its practical engineering applications, this paper has made systematic in depth studies on various aspects as preprocessing of IR images, target image segmentation , target recognition and target tracking. The main achievements of this paper are as fellows:
(1)The thermal imaging theory and the characteristics of infrared image are analyzed in this dissertation. Image histogram is mainly discussed and an interlace row/column histogram statistic method is proposed. Based on histogram research, two novel image enhancement algorithms are put forward. It is adaptive piecewise linear transform.
(2)A Wavelet edge detection method which is self-adaptive, multi -scale and anti-jamming has been studies for use in formation processing for IR imaging terminal homing guidance. By means of multi-resolution analysis concept for auto-zooming of wavelet transformation, a multi-resolution Gaussian gradient operator which is suitable for noisy edge detection is configured .The experiment results show that this method is an automatic rapid and effective target detection method.
(3)ATR is studied in this paper both invariant moments to round degree. And , against the helicopter the result of test is given. The idea of topological is introduced in the fried of ATR, the hitting-point choice on the basis of characteristic points is studied and aiming at the tanker a applied algorithm is proposed..
(4)It studies the target tracking, the hitting-point tracking is proposed .What's more, based on the comparison and analysis of the various algorithms of tracking and
target prediction, a new Kalman filter algorithm is presented.
(1)结合热成像技术原理对红外图像的特征进行了分析，重点研究了红外图像的灰度直方图，提出了隔行／列直方图统计的思想。以灰度直方图作为统计手段和依据，提出了自适应灰度分段线性变换的图像增强算法。
(2)研究了一种用于红外成像制导信息处理的自适应、多尺度、抗噪声小波边缘检测方法。该方法利用小波变换的自动变焦的多分辨分析思想，构造了适合的噪声中边缘检测的多分辨率高斯梯度算子。实验结果表明，该方法是一种快速、有效的自动化目标检测方法。
(3)在模式识别理论的基础上，探讨了图像识别系统，研究了不变量(不变矩、圆度)在目标识别中的应用，并针对圆度法用于直升机识别给出了应用实例，提出拓扑思想在识别中应用的方案，并针对坦克提出了一种命中点选择方案和算法。
(4)介绍了成像跟踪常用的两大方法，波门法、匹配法，并在波门法中提出了基于命中点选择的命中点跟踪法，研究了跟踪预测算法的应用，针对红外图像目标提出了一种卡尔曼滤波改进算法。
With the development of precision guided weapons, the requirements on information processing techniques for precision guidance is becoming higher and higher. This has become a key factor, which restricts the development of the weapon. Since IR imaging guidance possesses many advantages, it has become the trend and direction for current development and study of precision guidance technologies worldwide. On the basis of the fundamental characteristics of IR imaging guidance, the current development status of guidance information processing methods, the main problems existed as well as the basic requirements on its practical engineering applications, this paper has made systematic in depth studies on various aspects as preprocessing of IR images, target image segmentation , target recognition and target tracking. The main achievements of this paper are as fellows:
(1)The thermal imaging theory and the characteristics of infrared image are analyzed in this dissertation. Image histogram is mainly discussed and an interlace row/column histogram statistic method is proposed. Based on histogram research, two novel image enhancement algorithms are put forward. It is adaptive piecewise linear transform.
(2)A Wavelet edge detection method which is self-adaptive, multi -scale and anti-jamming has been studies for use in formation processing for IR imaging terminal homing guidance. By means of multi-resolution analysis concept for auto-zooming of wavelet transformation, a multi-resolution Gaussian gradient operator which is suitable for noisy edge detection is configured .The experiment results show that this method is an automatic rapid and effective target detection method.
(3)ATR is studied in this paper both invariant moments to round degree. And , against the helicopter the result of test is given. The idea of topological is introduced in the fried of ATR, the hitting-point choice on the basis of characteristic points is studied and aiming at the tanker a applied algorithm is proposed..
(4)It studies the target tracking, the hitting-point tracking is proposed .What's more, based on the comparison and analysis of the various algorithms of tracking and
target prediction, a new Kalman filter algorithm is presented.
引文
[1] 徐南荣，卞南华，红外辐射与制导，国防工业出版社，1997年6月。
[2] 王义玉，红外探测器，兵器工业出版社，1993年8月。
[3] 谭显裕，红外目标图像识别跟踪研究及其军用前景，光电子技术，1997年9月。
[4] 章毓晋编著，图像工程上册——图像处理和分析，清华大学出版社，1999年3月
[5] Kenneth．R．Castleman著，朱志刚，林学(门言)，石定机等译，数字图像处理，电子工业出版社，1998年9月
[6] 桑农，张天序，结合背景抑制技术的FLIR图像分割，红外与毫米波学报，Vol．18，No．5，Oct 1999，P．369-374
[7] 韩克松，陈景春，崔常嵬，红外图像中运动点目标检测技术，系统工程与电子技术，1998年第四期，P．23-26
[8] 王自勇，红外图像自动目标识别技术进展．飞航导弹情报研究报告文集(二)，中国航天三院第310所，1996，77～88
[9] 廖朝佩，战术导弹导引头技术进展．飞航导弹，1994，(1)：35～39
[10] 章毓晋，图像分割．北京：科学出版社，2001，67～70
[11] 金琪，戴汝为，矩、不变矩、角度矩及傅立叶描绘子．模式识别与人工智能，1993，6(3)：181～188
[12] 陈燕新，戚飞虎，一种新的提取轮廓特征点的方法．红外与毫米波学报．1998，17(3)：171～176
[13] 吴佩伦，精确制导武器的关键技术和战略发展方向．现代防御技术，1994，(5)：52～59
[14] 左震，张天序，汪国有，远距红外图像中桥梁目标识别方法研究，电子学报，1998，26(11)：6～9
[15] 卢福刚，雷万云，赵荣椿，复杂背景下红外图像目标自动检测，弹箭与制导学报，1998，(1)：18～23
[16] 何国威，刘建国，周春晓，张天序，基于知识的红外图像数字图像中铁路的识别，红外与激光工程，2000，29(6)：55～60
[17] 袁再华，杨树谦，桥梁红外图像的自动识别跟踪方法探讨，红外与激光工程，1998，27(1)：4～8
[18] 王栓，艾海舟，何克忠，基于差分图像的多运动目标的检测与跟踪，中国图象图形学报，1999，4(6)：470～474
[19] 王国宏，毛士艺，何友，红外传感器目标跟踪算法，火力与指挥控制，2001，26(2)：5～9
[20] 薛东辉，朱耀庭，朱光喜，宋琪，低对比度前视红外坦克目标图像检测方法研究，红外与毫米波学报，1996，15(3)：183～187
[21] 夏良正，数字图像处理，东南大学出版社，1999
[22] 崔锦泰，小波分析导论，西安交通大学出版社，1995
[23] 程正兴，小波分析算法与应用，西安交通大学出版社，1999
[24] 刘永昌，红外寻的制导技术的发展现状与新趋势，红外技术，vol．21 No 4
[25] 郑南宁，计算机视觉与模式识别，国防工业出版社，1998
[26] R．C．冈萨雷斯(美)，《数字图像处理》，中译本
[27] 边肇祺、张学工，模式识别，清华大学出版社，2000
[28] Bento Correia, Joao Dinis, Roger Davies. Automatic detection and recognitiong of stationary motorized vehicles in infrared images. SPIE Vol. 3718,1999
[29] Markus Müller, Axel Korn. Automatic Target Detection in Cluttered IR Images. SPIE Vol. 3375,1998.4
[30] D.J. Gregoris, S.K.W. Yu, S. Tritchew. Detection of dim targets in FLIR imagery using multiscale transforms. SPIE Vol. 2269,1994
[31] David Klick, Philip Blumenau, Joseph Theriault. Detection of targets in infrared clutter. Proceedings of SPIE VOL. 4370,2001
[32] Hankan Bakircioglu, Erol Gelenbe. Feature-based RNN Target Recognition. SPIE Vol 3370,1998.4
[33] J.A. Silva, An Adaptive Image Enhancement Algorithm, Pattern
Recognation, Vol. 30. No. 7,1997
[34] Li Biao, A Real-Time System for Target Detecting and Tracking in IR Image Sequence, 1996 IEEE Conference
[35] An Adaptive Image Segmentation Method with Visual Nonlinear Characteristics, IEEE Trans SMC, Vol .26, No. 8,1996
[36] 戚飞虎译，模式识别与图像处理，上海交通大学出版社，1989
[37] 杨煊等，基于边缘视觉特征描述的自适应多尺度小波边缘检测方法，电子科学学刊，Vol．22，No．6，1998
[38] 贾天旭等，基于Bubble小波的多尺度边缘提取，电子学报，Vol．24No．4．1996
[39] 雷万云，王立，李言俊，基于小波变换的红外成像末制导方法研究弹箭与制导学报
[40] D. Nair,J.K. Aggarwal. Bayesian reconition of targets by parts in second generation forward looking infrared images. Image and Vision Computing (18) ,2000
[41] E. Magli, G. Olmo, L. Lo Presti. Pattern recognition by means of the Radon transform and the continuous wavelet transform. Signal Processing 73 (1999) 277 -289,1998.12
[42] Jun Xu, Jiang-qi Zhang, Chang-hong Liang. Prediction of the performance an algorithm for the detectiong of small targets in infrared images, lnfrared Physics & Technology 42(2001) 17-22,2000.4
[43] ANIL K.JAIN, NALINI K. RATHA, SRIDHAR LAKSHMANAN. Object Detection using Gabor filters. Pattern Recognition, Vol. 30, NO. 2,295-309,1997
[44] Xiao-Ping Zhang, Mita O. Desai. Segmentation of Bright Targets Using Wavelets and Adaptive Thresholding. IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL 10, N0.7,2001.7
[45] E. Dror-Rein, H.B. Mitchell. A "soft" assignment algorithm for multiple target tracking .Sinal Processing 81,2001
[46] Malcolm J.M. Strens, Ian N. Gregory. Tracking in cluttered images. Image and Vision Computing, 2003.4
[47] Adrian G. Bors, loannis Pitas, Prediction and Tracking of Moving Objects in Image Sequences. IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL 9, N0.8,2000.8
[48] Sylvia Gil, Ruggero Milanese, Thierry Pun. Comparing features for Targets Tracking in Traffic Scenes. Pattern Recognition. Vol 29. N0.8,1996
[49] Cor J. Veenman, Marcel J.T. Reinders, Eric Backer. A Cellular Coevolutionary Algorithm for Image Segmentation. IEEE TRANSACTION TRANSACTIONS ON IMAGE PROCESSING, VOL 12, NO 3,2003.4
[50] M. Oiani, G. Corsini, A. Baldacci. Space-time processing for the
detection of airborne targets in IR image sequences. Image Signal Process. Vol 148, NO 3,2001.6
[51] 赵荣椿，数字图像处理导论，西北工业大学出版社，1995
[52] 吕风军，数字图像处理编程入门，清华大学出版社，1999
[53] 李宏贵，罗正发，李兴国，一种基于小波变换的红外图像识别方法．模式识别与人工智能，Vol 13，No 1，2000．4
[54] 刘其真，姚剑，孙薇，何永保，红外成像运动目标识别与跟踪方法研究．遥感技术与应用，Vol 14，No 2，1999．6
[55] Thomas Meitzler, Grant Gerhart, Euljung Sohn. Detection Probability Using Relative Clutter in Infrared Images. IEEE TRANSACTIONS ON AEROSPACE AND EECTRONIC SYSTEM, Vol. 34, NO. 3,1998.7
[56] Shan Cong, Lang Hong. Interval Models for Target Tracking Algorithms. Mathematical and Computer Modelling 34(2001)593-602
[57] 阎平凡等，模式分类器和可训练机器，科学出版社，1987
[58] 颜孙震、钟山等，矩不变量在目标形状识别中的应用研究，国防科技大学学报，1998 Vol．20．No．5
[59] 阎平凡、黄瑞旭，人工神经网络—模型、分析与应用，安徽教育出版社，1993
[2] 王义玉，红外探测器，兵器工业出版社，1993年8月。
[3] 谭显裕，红外目标图像识别跟踪研究及其军用前景，光电子技术，1997年9月。
[4] 章毓晋编著，图像工程上册——图像处理和分析，清华大学出版社，1999年3月
[5] Kenneth．R．Castleman著，朱志刚，林学(门言)，石定机等译，数字图像处理，电子工业出版社，1998年9月
[6] 桑农，张天序，结合背景抑制技术的FLIR图像分割，红外与毫米波学报，Vol．18，No．5，Oct 1999，P．369-374
[7] 韩克松，陈景春，崔常嵬，红外图像中运动点目标检测技术，系统工程与电子技术，1998年第四期，P．23-26
[8] 王自勇，红外图像自动目标识别技术进展．飞航导弹情报研究报告文集(二)，中国航天三院第310所，1996，77～88
[9] 廖朝佩，战术导弹导引头技术进展．飞航导弹，1994，(1)：35～39
[10] 章毓晋，图像分割．北京：科学出版社，2001，67～70
[11] 金琪，戴汝为，矩、不变矩、角度矩及傅立叶描绘子．模式识别与人工智能，1993，6(3)：181～188
[12] 陈燕新，戚飞虎，一种新的提取轮廓特征点的方法．红外与毫米波学报．1998，17(3)：171～176
[13] 吴佩伦，精确制导武器的关键技术和战略发展方向．现代防御技术，1994，(5)：52～59
[14] 左震，张天序，汪国有，远距红外图像中桥梁目标识别方法研究，电子学报，1998，26(11)：6～9
[15] 卢福刚，雷万云，赵荣椿，复杂背景下红外图像目标自动检测，弹箭与制导学报，1998，(1)：18～23
[16] 何国威，刘建国，周春晓，张天序，基于知识的红外图像数字图像中铁路的识别，红外与激光工程，2000，29(6)：55～60
[17] 袁再华，杨树谦，桥梁红外图像的自动识别跟踪方法探讨，红外与激光工程，1998，27(1)：4～8
[18] 王栓，艾海舟，何克忠，基于差分图像的多运动目标的检测与跟踪，中国图象图形学报，1999，4(6)：470～474
[19] 王国宏，毛士艺，何友，红外传感器目标跟踪算法，火力与指挥控制，2001，26(2)：5～9
[20] 薛东辉，朱耀庭，朱光喜，宋琪，低对比度前视红外坦克目标图像检测方法研究，红外与毫米波学报，1996，15(3)：183～187
[21] 夏良正，数字图像处理，东南大学出版社，1999
[22] 崔锦泰，小波分析导论，西安交通大学出版社，1995
[23] 程正兴，小波分析算法与应用，西安交通大学出版社，1999
[24] 刘永昌，红外寻的制导技术的发展现状与新趋势，红外技术，vol．21 No 4
[25] 郑南宁，计算机视觉与模式识别，国防工业出版社，1998
[26] R．C．冈萨雷斯(美)，《数字图像处理》，中译本
[27] 边肇祺、张学工，模式识别，清华大学出版社，2000
[28] Bento Correia, Joao Dinis, Roger Davies. Automatic detection and recognitiong of stationary motorized vehicles in infrared images. SPIE Vol. 3718,1999
[29] Markus Müller, Axel Korn. Automatic Target Detection in Cluttered IR Images. SPIE Vol. 3375,1998.4
[30] D.J. Gregoris, S.K.W. Yu, S. Tritchew. Detection of dim targets in FLIR imagery using multiscale transforms. SPIE Vol. 2269,1994
[31] David Klick, Philip Blumenau, Joseph Theriault. Detection of targets in infrared clutter. Proceedings of SPIE VOL. 4370,2001
[32] Hankan Bakircioglu, Erol Gelenbe. Feature-based RNN Target Recognition. SPIE Vol 3370,1998.4
[33] J.A. Silva, An Adaptive Image Enhancement Algorithm, Pattern
Recognation, Vol. 30. No. 7,1997
[34] Li Biao, A Real-Time System for Target Detecting and Tracking in IR Image Sequence, 1996 IEEE Conference
[35] An Adaptive Image Segmentation Method with Visual Nonlinear Characteristics, IEEE Trans SMC, Vol .26, No. 8,1996
[36] 戚飞虎译，模式识别与图像处理，上海交通大学出版社，1989
[37] 杨煊等，基于边缘视觉特征描述的自适应多尺度小波边缘检测方法，电子科学学刊，Vol．22，No．6，1998
[38] 贾天旭等，基于Bubble小波的多尺度边缘提取，电子学报，Vol．24No．4．1996
[39] 雷万云，王立，李言俊，基于小波变换的红外成像末制导方法研究弹箭与制导学报
[40] D. Nair,J.K. Aggarwal. Bayesian reconition of targets by parts in second generation forward looking infrared images. Image and Vision Computing (18) ,2000
[41] E. Magli, G. Olmo, L. Lo Presti. Pattern recognition by means of the Radon transform and the continuous wavelet transform. Signal Processing 73 (1999) 277 -289,1998.12
[42] Jun Xu, Jiang-qi Zhang, Chang-hong Liang. Prediction of the performance an algorithm for the detectiong of small targets in infrared images, lnfrared Physics & Technology 42(2001) 17-22,2000.4
[43] ANIL K.JAIN, NALINI K. RATHA, SRIDHAR LAKSHMANAN. Object Detection using Gabor filters. Pattern Recognition, Vol. 30, NO. 2,295-309,1997
[44] Xiao-Ping Zhang, Mita O. Desai. Segmentation of Bright Targets Using Wavelets and Adaptive Thresholding. IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL 10, N0.7,2001.7
[45] E. Dror-Rein, H.B. Mitchell. A "soft" assignment algorithm for multiple target tracking .Sinal Processing 81,2001
[46] Malcolm J.M. Strens, Ian N. Gregory. Tracking in cluttered images. Image and Vision Computing, 2003.4
[47] Adrian G. Bors, loannis Pitas, Prediction and Tracking of Moving Objects in Image Sequences. IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL 9, N0.8,2000.8
[48] Sylvia Gil, Ruggero Milanese, Thierry Pun. Comparing features for Targets Tracking in Traffic Scenes. Pattern Recognition. Vol 29. N0.8,1996
[49] Cor J. Veenman, Marcel J.T. Reinders, Eric Backer. A Cellular Coevolutionary Algorithm for Image Segmentation. IEEE TRANSACTION TRANSACTIONS ON IMAGE PROCESSING, VOL 12, NO 3,2003.4
[50] M. Oiani, G. Corsini, A. Baldacci. Space-time processing for the
detection of airborne targets in IR image sequences. Image Signal Process. Vol 148, NO 3,2001.6
[51] 赵荣椿，数字图像处理导论，西北工业大学出版社，1995
[52] 吕风军，数字图像处理编程入门，清华大学出版社，1999
[53] 李宏贵，罗正发，李兴国，一种基于小波变换的红外图像识别方法．模式识别与人工智能，Vol 13，No 1，2000．4
[54] 刘其真，姚剑，孙薇，何永保，红外成像运动目标识别与跟踪方法研究．遥感技术与应用，Vol 14，No 2，1999．6
[55] Thomas Meitzler, Grant Gerhart, Euljung Sohn. Detection Probability Using Relative Clutter in Infrared Images. IEEE TRANSACTIONS ON AEROSPACE AND EECTRONIC SYSTEM, Vol. 34, NO. 3,1998.7
[56] Shan Cong, Lang Hong. Interval Models for Target Tracking Algorithms. Mathematical and Computer Modelling 34(2001)593-602
[57] 阎平凡等，模式分类器和可训练机器，科学出版社，1987
[58] 颜孙震、钟山等，矩不变量在目标形状识别中的应用研究，国防科技大学学报，1998 Vol．20．No．5
[59] 阎平凡、黄瑞旭，人工神经网络—模型、分析与应用，安徽教育出版社，1993