基于光流分析的无人机视频运动目标检测与跟踪
|
摘要
运动目标的检测与跟踪是计算机视觉研究中的热点和难点之一,在智能交通、安全监控以及军事领域有着广阔的应用前景。无人机具有高机动性、高分辨率、隐蔽性好、操作灵活等优势,主要应用于昼夜空中侦察、战场监视、战场毁伤评估和军事测绘领域。利用无人机搭载的视频传感器对地面运动目标进行跟踪与分析,在军事上具有重要的实践意义和理论价值。然而由于视频传感器随着无人机的高速运动而运动以及视频序列影像中背景的复杂性和运动目标信息的多样性的特点,使得处理目标检测跟踪的问题变得更加困难。本文研究的目的是将光流技术引入到运动目标检测跟踪研究中,完成的主要工作及取得的成果如下:
1.阐述了光流技术的基本原理和常用的计算方法,在此基础上,分析并总结了光流技术在运动估计、目标检测与跟踪中的应用方案,为后续各章节利用光流技术的应用奠定了理论基础。
2.为了消除相机运动引起的影像背景的位移,运用了基于Harris特征点匹配的运动估计和全局运动补偿方法;提出了一种利用光流技术进行背景运动估计,建立背景运动参数模型,并进行全局运动补偿的方法。通过实验对比,证实了光流技术方法具有计算速度快,实用性强的优势。
3.简单分析了静止背景条件下运动目标的检测方法,实现了经过背景运动补偿后的简单差分运动目标检测。重点分析了用于轮廓分割的水平集法,并将光流技术和水平集相结合,实现了运动目标的检测和分割,在静止和运动背景条件下取得了比较满意的实验结果。
4.在动态背景下,通过光流方法进行背景运动估计、全局运动补偿以及运动目标的检测与分割,在此基础上,采用连续自适应均值漂移(Camshift)算法和卡尔曼滤波状态预测两种方法进行运动目标的跟踪,并实现了单目标和多目标的连续跟踪。
Moving target detection and object tracking, is one of the hot and difficult topics in computer vision, which has broad application prospects in intelligent transportation, security monitoring and other military fields. Unmanned Aerial vehicles (UAVs) which has advantages of high maneuverability, high resolution, convenient hiding and flexible operation, is mainly used in military surveying and mapping fields ,day and night aerial reconnaissance, battlefield surveillance, battle damage assessment and etc. Therefore, UAVs which is equipped with video sensors to track and analyze the moving ground target that has important practical significance and theoretical value in military area. However, because of the characteristics that the moving video sensor is moving with the high-speed UAVs simultaneously, the complexity of the background in the video image sequence and the diversity of moving target information, which makes the problem of treatment target detection and object tracking becomes even more difficult. The research purpose of this paper is to bring optical flow technology into the moving target detection and tracking field. The major work and achieved results now has been completed are as follows:
1. Describe the basic principles of optical flow and common calculation methods, and upon this, analyze and summarize the application plan that how the optical flow technology is used in motion estimation and object detection and tracking, which laid a theoretical foundation for the following chapters of the application of optical flow technology.
2. In order to eliminate the displacement of the image background which is caused by camera movement , this paper uses motion estimation and global motion compensation,which is based on the harris feature points matching, to proposes a method which uses optical flow technology to calculate background motion estimation, build background motion parameter model, and finally compensate the global motion; Comparing experiments confirms that the optical flow method have great advantages of rapid calculating and great practicability.
3. Moving target detection method is briefly analyzed and difference moving target detection method is realized after background motion compensation. The content of this chapter emphasizes on the level method that is used in contour segmentation and combines the optical flow and level set to facilitate moving target detection and segmentation, which achieves satisfying results in both still and moving background conditions .
4. Background motion estimation, global motion compensation and the detection and segmentation of moving target is realized through the optical flow method in dynamic background. And based on the method, this chapter uses continuous adaptive mean shift(Camshift) and Kalman filter prediction method to track target and thereby, achieves continuous tracking of both single and mutiple targets.
1.阐述了光流技术的基本原理和常用的计算方法,在此基础上,分析并总结了光流技术在运动估计、目标检测与跟踪中的应用方案,为后续各章节利用光流技术的应用奠定了理论基础。
2.为了消除相机运动引起的影像背景的位移,运用了基于Harris特征点匹配的运动估计和全局运动补偿方法;提出了一种利用光流技术进行背景运动估计,建立背景运动参数模型,并进行全局运动补偿的方法。通过实验对比,证实了光流技术方法具有计算速度快,实用性强的优势。
3.简单分析了静止背景条件下运动目标的检测方法,实现了经过背景运动补偿后的简单差分运动目标检测。重点分析了用于轮廓分割的水平集法,并将光流技术和水平集相结合,实现了运动目标的检测和分割,在静止和运动背景条件下取得了比较满意的实验结果。
4.在动态背景下,通过光流方法进行背景运动估计、全局运动补偿以及运动目标的检测与分割,在此基础上,采用连续自适应均值漂移(Camshift)算法和卡尔曼滤波状态预测两种方法进行运动目标的跟踪,并实现了单目标和多目标的连续跟踪。
Moving target detection and object tracking, is one of the hot and difficult topics in computer vision, which has broad application prospects in intelligent transportation, security monitoring and other military fields. Unmanned Aerial vehicles (UAVs) which has advantages of high maneuverability, high resolution, convenient hiding and flexible operation, is mainly used in military surveying and mapping fields ,day and night aerial reconnaissance, battlefield surveillance, battle damage assessment and etc. Therefore, UAVs which is equipped with video sensors to track and analyze the moving ground target that has important practical significance and theoretical value in military area. However, because of the characteristics that the moving video sensor is moving with the high-speed UAVs simultaneously, the complexity of the background in the video image sequence and the diversity of moving target information, which makes the problem of treatment target detection and object tracking becomes even more difficult. The research purpose of this paper is to bring optical flow technology into the moving target detection and tracking field. The major work and achieved results now has been completed are as follows:
1. Describe the basic principles of optical flow and common calculation methods, and upon this, analyze and summarize the application plan that how the optical flow technology is used in motion estimation and object detection and tracking, which laid a theoretical foundation for the following chapters of the application of optical flow technology.
2. In order to eliminate the displacement of the image background which is caused by camera movement , this paper uses motion estimation and global motion compensation,which is based on the harris feature points matching, to proposes a method which uses optical flow technology to calculate background motion estimation, build background motion parameter model, and finally compensate the global motion; Comparing experiments confirms that the optical flow method have great advantages of rapid calculating and great practicability.
3. Moving target detection method is briefly analyzed and difference moving target detection method is realized after background motion compensation. The content of this chapter emphasizes on the level method that is used in contour segmentation and combines the optical flow and level set to facilitate moving target detection and segmentation, which achieves satisfying results in both still and moving background conditions .
4. Background motion estimation, global motion compensation and the detection and segmentation of moving target is realized through the optical flow method in dynamic background. And based on the method, this chapter uses continuous adaptive mean shift(Camshift) and Kalman filter prediction method to track target and thereby, achieves continuous tracking of both single and mutiple targets.
引文
[1]苏鑫鑫,王永寿.美军无人机系统研发现状与动向[J].飞航导弹,2009.12.
[2]丁土.中国无人机发展之路[J].无人战机探秘.2007.09
[3]林一平.我国高校无人机研发成果丰硕[J].航空航天.2009.02
[4] Robert T. Collins, Alan J. Lipton, Takeo Kanade etc. A System for Video Surveillance and Monitoring[R]. CMU-RI-TR-00-12, Carnegie Mellon University.
[5] DTO.Data Mining Report[R].Office of the Director of Nation Intelligence,American,2009
[6] Saad Ali and Mubarak Shah. COCOA - Tracking in Aerial Imagery[J]. SPIE Airborne Intelligence, Surveillance, Reconnaissance (ISR) Systems and Applications, Orlando, 2006.
[7]贾云德.机器视觉[M],科学出版社,北京,2000.4
[8] D.J.Fleet, and A.D.Jepson,Computation of component image velocity from local phase information ,IJCV5(1)(1990),pp77-104
[9] D.J.Heeger,Optical flow using spatiotemporal ,IJCV,1(1988),pp279-302
[10] Jain R,Nagel H.On the analysis of accumulative difference of picture from image sequences of real world scenes[J].IEEE Trans.PAMI,1979:206~214
[11]陈朝阳,张桂林.基于图像对称差分运算的运动小目标检测方法[J].华中理工大学学报,1998,26(9):34~38
[12] Lipton A,Fujiyoshi H and Patil R.Moving target classification and tracking from real time video[C].Proc of the 1998 DARPA Image Understanding Workshop,1998
[13] Changick K.Fast and automatic video object segmentation and tracking for content bassed application[J].IEEE Transaction on Circuits and System for Video Technology,2002,12(2):122~129
[14]杨志华,曾禹村.背景移动补偿技术的研究[J].北京理工大学学报,2000,20(3):333~337
[15] Araki S,Matsuoaka T,Yokoya N,Takemura H.Real-time tracking of multiple moving object contours in a moving camera image sequence[J].IEICE Trans. Inf.&Syst.,2000,(7):131~137
[16] Eun J H,Robyn O.Segmenting occluded objects using a motion snake[C].Asi Conference on Computer Vision,2004
[17] Alexei A E,Alexander C B,Greg M,Jitendra M.Recognizing Action at distance[C].International Conference on Computer Vision,2003:560~566
[18] Mae Y,Shirai Y,Miura J,et al.Object tracking in cluttered background based optical flow and edges[C].Proceedings of the 13th International Conference Pattern Recognition,1996,1:196~200
[19]张磊,项学智,赵春晖.基于光流场与水平集的运动目标检测[J].计算机应用.2009.4
[20]黄赞,张宪民.改进的基于光流的鲁棒多尺度运动估计算法[J].华南理工大学学报(自然科学版).2009.12
[21] Ahmed E,David H,Larry D.Non-parametric Model for background subtraction[C].European Conference on Computer Vision,2000:654~659
[22] Henry S and Takeo K.Object detection using the statistics of parts[J].International Journal of ComputerVision,2004,56(3):151~177
[23] WrenC.R.,AzarbayejaniA.,etal.Pfinder:real一time traeking of the human body,IEEE Transalationson Pattern Analysis and Maehine Intelligenee.1997,19(7):780-785
[24] JPL.Traffic surveillance and deteetion technology development.Senfor Development Final ReP.97-10,Jet ProPulsion Laboratory,1997
[25] Kass M.,Witkin A.,Terzopoulous D.Snakes:Active contour models[J].International Journal of Computer Vision,1987,1(4):321-331.
[26] Nickels K.,Hutchinson S.Estimating uncertainty in SSD-based feature tracking[J].Image and Vision Computing.2002,20(1):47-58.
[27] Zhu Z.,Xu G.,Yang B.,Shi D.,Lin X.VISATRAM:A real-time vision system for automatic traffic monitoring[J].Image and Vision Computing,2000,18(10):781-794.
[28] Berthold K.P.Horn ,Bian G..Schunck.Determing Optical Flow[J]. artificial intelligence,1981
[29] Barron,Beauchemin,Fleet.On Optical Flow[J]. AIICSR,1994.9
[30]董颖.基于光流场的视频运动检测[D].山东大学硕士学位论文.2008.05
[31] Shijun Sun, David Haynor, and Yongmin Kim. Motion Estimation Based on Optical Flow with Adaptive Gradients[J].IEEE,2000
[32]武理静,明军.一种改进的光流计算方法[J].国外电子测量技术,2006(12)
[33] Shinya Yamamoto, Yasushi Mae. Realtime Multiple Object Tracking Based on Optical Flows [J]. IEEE International Conference on Robotics and Automation,1995
[34]雷震,吴玲达,李东.基于局部微分光流的运动对象分割[J].计算机科学.2009.6
[35] Adiv G D.Determing 3-D motion and structure from optical flow generated by several moving objects[J].IEEE Transactions on Pattern Analysis and machine Intelligence,1985
[36]罗寰,廖俊,穆忠林,于雷.基于光流估计得红外弱小运动目标检测[J].电光与控制.2009.07
[37] Harris C,Stephens M. A Combined Corner and Edge Detector [C]. Fourth Alvey Vision Conference,Manchester UK,1988:147-151
[38]吴福朝.计算机视觉中的数学方法[M],科学出版社,北京,2008.3
[39] ANDR′ES BRUHN, JOACHIM WEICKERT. Lucas/Kanade Meets Horn/Schunck: Combining Local and Global Optic Flow Methods. International Journal of Computer Vision 61(3), 211–231, 2005
[40] V.Markandey.Motion estimation for moving target deteetion. IEEE Transon AES, 1996, 32(3) : 866-874
[41]陈祖爵,陈潇君,何鸿.基于改进的混合高斯模型的运动目标检测[J],中国图象图形学报.2007.9
[42]刘光宇,刘国栋.基于混合高斯模型的运动目标检测方法[J],计算机工程与应用.2009
[43]刘英霞,贺长伟,王欣.基于贝叶斯模型的动态背景检测[J].系统仿真学报.2007.11
[44]李乃安.视频序列中运动目标检测与跟踪技术研究[D].桂林电子科技大学硕士论文.2008.
[45] Michael J. Black, Allan Jepson. Estimating Optical Flow in Segmented Images using Variable-order Parametric Models with Local Deformations [J]. IEEE PAMI.1996
[46] OSher S,Sethian J.Fronts propagating with curvature dependent speed :Algorithms based on the Hamilton-Jacobi formulation[J]. Jounal of Computation Physics,1988
[47]于慧敏,许艺,刘继忠,高晓颖.基于水平集的多运动目标时空分割与跟踪[J].中国图像图形学报.2007.07
[48]于慧敏,尤育赛.基于水平集的多运动目标检测与分割[J].浙江大学学报(工学版).2007.3
[49]王大凯,侯榆青,彭进业.图像处理的偏微分方程方法[M],科学出版社,北京,2008.6
[50] J.A.Sethian.Level Set Methods and Fast marching Methods[D].Univeisity of California,Berkeley.1999
[51] SASA G, LONGCARIC S. Spatio-temporal image segmentation using optical flow and lustering[C].Processing of the 1st International workshop on image and Signal processing and analysis:IW ISPA 2000.Pula,Croatia.2000
[52] Jeongho Shin, Sangjin Kim, Sangkyu Kang etc. Optical flow-based real-time object tracking using non-prior training active feature mode, ELSEVIER Real-Time Imaging, Vol. 11, pp. 204-218, USA, June 2005.
[53] Stanley Osher,Ronald Fedkiw. Level Set Methods and Dynamic Implicit Surfaces[M].New York:Springer Press 2002
[54] Whitaker R T.A level set approach to 3 D reconstruction from range data[J].The International Journal of Computer Vision,1998
[55]于文率.无人飞行器序列影像运动分析关键技术研究[D].解放军信息工程大学博士学位论文.2009.10
[56]李培华.序列图像中运动目标跟踪方法[M],科学出版社,北京,2010.4
[57]刘雪,常发亮,王华杰.基于改进Camshift算法的视频对象跟踪方法[J].图像处理.2007
[58]刘振华.运动场景中目标检测与跟踪技术研究[D].南京航空航天大学硕士学位论文.2008.01
[59]孙季丰,王成清.基于特征光流和卡尔曼滤波的运动车辆跟踪[J].华南理工大学学报(自然科学版).2005.10
[60] Feghali R,Mitiche A.Spatiotemporal motion boundary detection and motion boundary velocity estimation for tracking moving objects with amoving camera:a level sets PDEs approach with concurrent camera motion compensation[J].IEEE Transactions on Image Processing,2004.
[61] J.A.Sethian,Level Set Methods and Fast Marching Methods[D].Univercity of California , Berkeley.1999
[62] R.Malladi and J.A.Sethian.Level Set and Fast Marching Methods in Image Processing and Computer vision[J].IEEE Transaction on Image Processing,1996
[63] Shiya Yamamoto,Yasushi Mae,Yoshiaki Shirai,Jun Miura.Realtime Multiple Object Tracking Based On Optical flows[J].IEEE International Conference on Robotics and Automation.1995
[64] Toby P.Breckon,Stuart E.Barnes,Marcin L.Eichner and Ken Wahren. Autonomous Real-Time Vehicle Detection from a Medium-Level UAV[J], IEEE Transaction on Image Processing,2008
[65] David A.Semko.Optical Flow Analysis and kalman Filter Tracking in Video Surveillance Algorithms[D].Naval Postgraduate School.2007.7
[66] Jincheol Ha, Christopher Alvino,Gallagher Pryor. Active Contours and Optical Flow for Automatic Tracking of Flying Vehicles[J]. Proceeding of the 2004 American Control Conference.2004.07
[67] A. Hilsmann and P. Eisert,“Deformable object tracking using optical flow constraints,”in Proc. 4th Int. Conf. on Visual Media Production (CVMP 2007), Nov. 2007
[68] Wei Niu, Long Jiao, Dan Han and Yuan-Fang Wang,“Real-Time Multi-person Tracking in Video Surveillance,”Proceedings of the 4th International Conf. on Information, Communications and Signal Processing, Singapore, December 2003.
[69] Weiming Hu, Tieniu Tan, Liang Wang and Steve Maybank,“A survey on Visual Surveillance of Object Motion and Behaviors,”IEEE Trans. On Systems, Man, and Cybernetics,2004
[70] Greg Welch, Gary Bishop.An Introduction to the Kalman Filter. Special Interest Group on Computer Graphics.Course 8.
[71] Isaac Cohen,G′erard Medioni. Detection and Tracking of Objects in Airborne Video Imagery[J]. CVPR’98Workshop on Interpretation of Visual Motion.1998
[72] Isaac Cohen,G′erard Medioni. Detecting and Tracking Moving Objects for Video Surveillance[J]. IEEE Proc. Computer Vision and Pattern Recognition. Jun. 23-25, 1999.
[73] Chunming Li , Chenyang Xu etc. Level Set Evolution Without Re-initialization: A New Variational Formulation[J]. Proceedings of the 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition.2005
[74] Gray Bradski and Adrian Kaebler.Learing OpenCV[M],2009.09
[75] Mae Y,Yamamoto S,Shirai Y,Miura Jun.Optical flow based realtime object tracking by active vision system[C].Proc.2nd Japan-France Congress Mechatronics,1994,2:545~548
[2]丁土.中国无人机发展之路[J].无人战机探秘.2007.09
[3]林一平.我国高校无人机研发成果丰硕[J].航空航天.2009.02
[4] Robert T. Collins, Alan J. Lipton, Takeo Kanade etc. A System for Video Surveillance and Monitoring[R]. CMU-RI-TR-00-12, Carnegie Mellon University.
[5] DTO.Data Mining Report[R].Office of the Director of Nation Intelligence,American,2009
[6] Saad Ali and Mubarak Shah. COCOA - Tracking in Aerial Imagery[J]. SPIE Airborne Intelligence, Surveillance, Reconnaissance (ISR) Systems and Applications, Orlando, 2006.
[7]贾云德.机器视觉[M],科学出版社,北京,2000.4
[8] D.J.Fleet, and A.D.Jepson,Computation of component image velocity from local phase information ,IJCV5(1)(1990),pp77-104
[9] D.J.Heeger,Optical flow using spatiotemporal ,IJCV,1(1988),pp279-302
[10] Jain R,Nagel H.On the analysis of accumulative difference of picture from image sequences of real world scenes[J].IEEE Trans.PAMI,1979:206~214
[11]陈朝阳,张桂林.基于图像对称差分运算的运动小目标检测方法[J].华中理工大学学报,1998,26(9):34~38
[12] Lipton A,Fujiyoshi H and Patil R.Moving target classification and tracking from real time video[C].Proc of the 1998 DARPA Image Understanding Workshop,1998
[13] Changick K.Fast and automatic video object segmentation and tracking for content bassed application[J].IEEE Transaction on Circuits and System for Video Technology,2002,12(2):122~129
[14]杨志华,曾禹村.背景移动补偿技术的研究[J].北京理工大学学报,2000,20(3):333~337
[15] Araki S,Matsuoaka T,Yokoya N,Takemura H.Real-time tracking of multiple moving object contours in a moving camera image sequence[J].IEICE Trans. Inf.&Syst.,2000,(7):131~137
[16] Eun J H,Robyn O.Segmenting occluded objects using a motion snake[C].Asi Conference on Computer Vision,2004
[17] Alexei A E,Alexander C B,Greg M,Jitendra M.Recognizing Action at distance[C].International Conference on Computer Vision,2003:560~566
[18] Mae Y,Shirai Y,Miura J,et al.Object tracking in cluttered background based optical flow and edges[C].Proceedings of the 13th International Conference Pattern Recognition,1996,1:196~200
[19]张磊,项学智,赵春晖.基于光流场与水平集的运动目标检测[J].计算机应用.2009.4
[20]黄赞,张宪民.改进的基于光流的鲁棒多尺度运动估计算法[J].华南理工大学学报(自然科学版).2009.12
[21] Ahmed E,David H,Larry D.Non-parametric Model for background subtraction[C].European Conference on Computer Vision,2000:654~659
[22] Henry S and Takeo K.Object detection using the statistics of parts[J].International Journal of ComputerVision,2004,56(3):151~177
[23] WrenC.R.,AzarbayejaniA.,etal.Pfinder:real一time traeking of the human body,IEEE Transalationson Pattern Analysis and Maehine Intelligenee.1997,19(7):780-785
[24] JPL.Traffic surveillance and deteetion technology development.Senfor Development Final ReP.97-10,Jet ProPulsion Laboratory,1997
[25] Kass M.,Witkin A.,Terzopoulous D.Snakes:Active contour models[J].International Journal of Computer Vision,1987,1(4):321-331.
[26] Nickels K.,Hutchinson S.Estimating uncertainty in SSD-based feature tracking[J].Image and Vision Computing.2002,20(1):47-58.
[27] Zhu Z.,Xu G.,Yang B.,Shi D.,Lin X.VISATRAM:A real-time vision system for automatic traffic monitoring[J].Image and Vision Computing,2000,18(10):781-794.
[28] Berthold K.P.Horn ,Bian G..Schunck.Determing Optical Flow[J]. artificial intelligence,1981
[29] Barron,Beauchemin,Fleet.On Optical Flow[J]. AIICSR,1994.9
[30]董颖.基于光流场的视频运动检测[D].山东大学硕士学位论文.2008.05
[31] Shijun Sun, David Haynor, and Yongmin Kim. Motion Estimation Based on Optical Flow with Adaptive Gradients[J].IEEE,2000
[32]武理静,明军.一种改进的光流计算方法[J].国外电子测量技术,2006(12)
[33] Shinya Yamamoto, Yasushi Mae. Realtime Multiple Object Tracking Based on Optical Flows [J]. IEEE International Conference on Robotics and Automation,1995
[34]雷震,吴玲达,李东.基于局部微分光流的运动对象分割[J].计算机科学.2009.6
[35] Adiv G D.Determing 3-D motion and structure from optical flow generated by several moving objects[J].IEEE Transactions on Pattern Analysis and machine Intelligence,1985
[36]罗寰,廖俊,穆忠林,于雷.基于光流估计得红外弱小运动目标检测[J].电光与控制.2009.07
[37] Harris C,Stephens M. A Combined Corner and Edge Detector [C]. Fourth Alvey Vision Conference,Manchester UK,1988:147-151
[38]吴福朝.计算机视觉中的数学方法[M],科学出版社,北京,2008.3
[39] ANDR′ES BRUHN, JOACHIM WEICKERT. Lucas/Kanade Meets Horn/Schunck: Combining Local and Global Optic Flow Methods. International Journal of Computer Vision 61(3), 211–231, 2005
[40] V.Markandey.Motion estimation for moving target deteetion. IEEE Transon AES, 1996, 32(3) : 866-874
[41]陈祖爵,陈潇君,何鸿.基于改进的混合高斯模型的运动目标检测[J],中国图象图形学报.2007.9
[42]刘光宇,刘国栋.基于混合高斯模型的运动目标检测方法[J],计算机工程与应用.2009
[43]刘英霞,贺长伟,王欣.基于贝叶斯模型的动态背景检测[J].系统仿真学报.2007.11
[44]李乃安.视频序列中运动目标检测与跟踪技术研究[D].桂林电子科技大学硕士论文.2008.
[45] Michael J. Black, Allan Jepson. Estimating Optical Flow in Segmented Images using Variable-order Parametric Models with Local Deformations [J]. IEEE PAMI.1996
[46] OSher S,Sethian J.Fronts propagating with curvature dependent speed :Algorithms based on the Hamilton-Jacobi formulation[J]. Jounal of Computation Physics,1988
[47]于慧敏,许艺,刘继忠,高晓颖.基于水平集的多运动目标时空分割与跟踪[J].中国图像图形学报.2007.07
[48]于慧敏,尤育赛.基于水平集的多运动目标检测与分割[J].浙江大学学报(工学版).2007.3
[49]王大凯,侯榆青,彭进业.图像处理的偏微分方程方法[M],科学出版社,北京,2008.6
[50] J.A.Sethian.Level Set Methods and Fast marching Methods[D].Univeisity of California,Berkeley.1999
[51] SASA G, LONGCARIC S. Spatio-temporal image segmentation using optical flow and lustering[C].Processing of the 1st International workshop on image and Signal processing and analysis:IW ISPA 2000.Pula,Croatia.2000
[52] Jeongho Shin, Sangjin Kim, Sangkyu Kang etc. Optical flow-based real-time object tracking using non-prior training active feature mode, ELSEVIER Real-Time Imaging, Vol. 11, pp. 204-218, USA, June 2005.
[53] Stanley Osher,Ronald Fedkiw. Level Set Methods and Dynamic Implicit Surfaces[M].New York:Springer Press 2002
[54] Whitaker R T.A level set approach to 3 D reconstruction from range data[J].The International Journal of Computer Vision,1998
[55]于文率.无人飞行器序列影像运动分析关键技术研究[D].解放军信息工程大学博士学位论文.2009.10
[56]李培华.序列图像中运动目标跟踪方法[M],科学出版社,北京,2010.4
[57]刘雪,常发亮,王华杰.基于改进Camshift算法的视频对象跟踪方法[J].图像处理.2007
[58]刘振华.运动场景中目标检测与跟踪技术研究[D].南京航空航天大学硕士学位论文.2008.01
[59]孙季丰,王成清.基于特征光流和卡尔曼滤波的运动车辆跟踪[J].华南理工大学学报(自然科学版).2005.10
[60] Feghali R,Mitiche A.Spatiotemporal motion boundary detection and motion boundary velocity estimation for tracking moving objects with amoving camera:a level sets PDEs approach with concurrent camera motion compensation[J].IEEE Transactions on Image Processing,2004.
[61] J.A.Sethian,Level Set Methods and Fast Marching Methods[D].Univercity of California , Berkeley.1999
[62] R.Malladi and J.A.Sethian.Level Set and Fast Marching Methods in Image Processing and Computer vision[J].IEEE Transaction on Image Processing,1996
[63] Shiya Yamamoto,Yasushi Mae,Yoshiaki Shirai,Jun Miura.Realtime Multiple Object Tracking Based On Optical flows[J].IEEE International Conference on Robotics and Automation.1995
[64] Toby P.Breckon,Stuart E.Barnes,Marcin L.Eichner and Ken Wahren. Autonomous Real-Time Vehicle Detection from a Medium-Level UAV[J], IEEE Transaction on Image Processing,2008
[65] David A.Semko.Optical Flow Analysis and kalman Filter Tracking in Video Surveillance Algorithms[D].Naval Postgraduate School.2007.7
[66] Jincheol Ha, Christopher Alvino,Gallagher Pryor. Active Contours and Optical Flow for Automatic Tracking of Flying Vehicles[J]. Proceeding of the 2004 American Control Conference.2004.07
[67] A. Hilsmann and P. Eisert,“Deformable object tracking using optical flow constraints,”in Proc. 4th Int. Conf. on Visual Media Production (CVMP 2007), Nov. 2007
[68] Wei Niu, Long Jiao, Dan Han and Yuan-Fang Wang,“Real-Time Multi-person Tracking in Video Surveillance,”Proceedings of the 4th International Conf. on Information, Communications and Signal Processing, Singapore, December 2003.
[69] Weiming Hu, Tieniu Tan, Liang Wang and Steve Maybank,“A survey on Visual Surveillance of Object Motion and Behaviors,”IEEE Trans. On Systems, Man, and Cybernetics,2004
[70] Greg Welch, Gary Bishop.An Introduction to the Kalman Filter. Special Interest Group on Computer Graphics.Course 8.
[71] Isaac Cohen,G′erard Medioni. Detection and Tracking of Objects in Airborne Video Imagery[J]. CVPR’98Workshop on Interpretation of Visual Motion.1998
[72] Isaac Cohen,G′erard Medioni. Detecting and Tracking Moving Objects for Video Surveillance[J]. IEEE Proc. Computer Vision and Pattern Recognition. Jun. 23-25, 1999.
[73] Chunming Li , Chenyang Xu etc. Level Set Evolution Without Re-initialization: A New Variational Formulation[J]. Proceedings of the 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition.2005
[74] Gray Bradski and Adrian Kaebler.Learing OpenCV[M],2009.09
[75] Mae Y,Yamamoto S,Shirai Y,Miura Jun.Optical flow based realtime object tracking by active vision system[C].Proc.2nd Japan-France Congress Mechatronics,1994,2:545~548