基于粒子滤波跟踪的步态特征提取算法研究
|
摘要
所谓步态,就是指人们行走的姿势,是一种远距离情况下唯一可感知的生物行为特征。它可以通过人步行的方式,达到识别人的身份的目的,近年来,成为国内外研究的重点领域。
步态识别是一种新兴的基于步态特征的生物特征识别技术,包括身高、体形等,它主要是针对含有人的运动图像序列进行分析处理,通常包括目标分割、特征提取、特征处理和识别分类四个阶段。它在虚拟现实,视觉监控,感知接口中均有广泛的应用前景。步态特征提取是步态识别中的一个重要步骤,现阶段,步态特征已经包含很多,例如：具有尺度、平移和旋转不变性,可以表示人体轮廓特征的傅里叶描述子；可以精确地表示不同人的身份,变化包含了大量的运动信息的肢体角度特征；表示步态对称性的反射对称因子；轮廓形状上下文等等。本文验证的是人体的傅里叶描述子特征。
智能视觉监控技术的研究近年来也受到了广泛的关注,即赋予监控系统观察分析场景内容的能力,使其更加智能化,能够在几乎不需要人为干预的情况下,对摄像机拍录的视频序列进行自动分析,并及时做出反应。例如对用户定义的异常情况的检测,或者一些非常规事件都能起到很好的作用。
运动目标的跟踪过程,就是依据目标及其所处的环境,选择能够唯一表示目标的特征,并在后续帧中搜索与该特征最匹配的目标位置的过程。它不仅是目标运动分析和场景分析的主要数据来源,也为目标的检测和识别提供帮助。目前关于目标跟踪(object tracking)的算法有很多,各种算法的根本区别在于选择何种特征对目标进行描述以及采用何种搜索匹配算法。本文采用的是粒子滤波算法。
运用两种技术的组合,实现对步态识别率的提高是本文的主要内容和创新,提出的方法是：首先运用背景减除方法从背景中提取出目标,并计算其傅里叶描述子特征,而后利用粒子滤波算法得到运动过程中的更为准确的目标位置信息,并在此基础上计算傅里叶描述子,实现更优越的步态识别效果。通过对滤波前后的傅里叶特征与真实目标特征的对比,证明经跟踪之后提取的特征信息更为准确,能够提高对人体目标的识别效率。
The so-called gait, which is the walking posture of people, is a biological behavioral feature which can be perceived only in the case of long-distance. It can achieve the purpose of recognising the identification of persons by the way people walk. It has become the focus of the study area at home and abroad recently.
Gait recognition is a new biometrics recognition technology based on gait characteristics, including height, body shape and so on.It is mainly aimed at analysis and processing of moving image sequence including person. Generally it includes four stages:target segmentation, feature extraction, feature processing and recognition classification. It has a broad application prospects in virtual reality, visual surveillance, perceptual interface. Gait feature extraction is an important step in gait recognition. Gait characteristics already contains a lot at present, for example:Fourier descriptor with the scale, translation and rotation invariance which can represent the profile characteristics of the human body; limb angle containing a great deal of motion information which can express the identity of different people accurately; Reflective symmetry factor that can represent gait symmetry; contour shape context and so on. This paper verifises Fourier descriptor of the human body.
The research of Intelligent Visual Surveillance has also been of widespread concern in recent years namely gives the monitoring system the ability to observe and analyze the content of the scene in order to make it more intelligent. It can analyze automatically the video sequence the camera records almost without human intervention, and timely respond. For example, it plays a very good role in the case of anomaly detection user-defined, or some unusual event.
Moving target tracking process is a process by selecting the only characteristics of the target and searching the target location which matches the characteristics in the follow-up frame based on the objectives and their environment. It is not only the main data source of target motion and scene analysis, but also of help for target detection and recognition. At present there are many algorithms on the target tracking (object tracking).Fundamental difference among the various algorithms is the choice of the characteristics of the target description and the type of search matching algorithm. In this paper, we introduce the particle filter algorithm.
The use of a combination of two techniques to achieve the improvement of the gait recognition rate is the main content and innovation of this article, the proposed method is:First, we use background subtraction method to extract the target from the background, and its Fourier descriptor features and then to use a particle filter algorithm to get a more accurate target location information in the process of movement, and on this basis, calculate Fourier descriptors to achieve superior gait recognition effect. It proves that the information extracted features with track is more accurate by the contrast of Fourier descriptor before and after filtering and the real target. This algorithm can improve the efficiency of target recognition of the human body.
步态识别是一种新兴的基于步态特征的生物特征识别技术,包括身高、体形等,它主要是针对含有人的运动图像序列进行分析处理,通常包括目标分割、特征提取、特征处理和识别分类四个阶段。它在虚拟现实,视觉监控,感知接口中均有广泛的应用前景。步态特征提取是步态识别中的一个重要步骤,现阶段,步态特征已经包含很多,例如：具有尺度、平移和旋转不变性,可以表示人体轮廓特征的傅里叶描述子；可以精确地表示不同人的身份,变化包含了大量的运动信息的肢体角度特征；表示步态对称性的反射对称因子；轮廓形状上下文等等。本文验证的是人体的傅里叶描述子特征。
智能视觉监控技术的研究近年来也受到了广泛的关注,即赋予监控系统观察分析场景内容的能力,使其更加智能化,能够在几乎不需要人为干预的情况下,对摄像机拍录的视频序列进行自动分析,并及时做出反应。例如对用户定义的异常情况的检测,或者一些非常规事件都能起到很好的作用。
运动目标的跟踪过程,就是依据目标及其所处的环境,选择能够唯一表示目标的特征,并在后续帧中搜索与该特征最匹配的目标位置的过程。它不仅是目标运动分析和场景分析的主要数据来源,也为目标的检测和识别提供帮助。目前关于目标跟踪(object tracking)的算法有很多,各种算法的根本区别在于选择何种特征对目标进行描述以及采用何种搜索匹配算法。本文采用的是粒子滤波算法。
运用两种技术的组合,实现对步态识别率的提高是本文的主要内容和创新,提出的方法是：首先运用背景减除方法从背景中提取出目标,并计算其傅里叶描述子特征,而后利用粒子滤波算法得到运动过程中的更为准确的目标位置信息,并在此基础上计算傅里叶描述子,实现更优越的步态识别效果。通过对滤波前后的傅里叶特征与真实目标特征的对比,证明经跟踪之后提取的特征信息更为准确,能够提高对人体目标的识别效率。
The so-called gait, which is the walking posture of people, is a biological behavioral feature which can be perceived only in the case of long-distance. It can achieve the purpose of recognising the identification of persons by the way people walk. It has become the focus of the study area at home and abroad recently.
Gait recognition is a new biometrics recognition technology based on gait characteristics, including height, body shape and so on.It is mainly aimed at analysis and processing of moving image sequence including person. Generally it includes four stages:target segmentation, feature extraction, feature processing and recognition classification. It has a broad application prospects in virtual reality, visual surveillance, perceptual interface. Gait feature extraction is an important step in gait recognition. Gait characteristics already contains a lot at present, for example:Fourier descriptor with the scale, translation and rotation invariance which can represent the profile characteristics of the human body; limb angle containing a great deal of motion information which can express the identity of different people accurately; Reflective symmetry factor that can represent gait symmetry; contour shape context and so on. This paper verifises Fourier descriptor of the human body.
The research of Intelligent Visual Surveillance has also been of widespread concern in recent years namely gives the monitoring system the ability to observe and analyze the content of the scene in order to make it more intelligent. It can analyze automatically the video sequence the camera records almost without human intervention, and timely respond. For example, it plays a very good role in the case of anomaly detection user-defined, or some unusual event.
Moving target tracking process is a process by selecting the only characteristics of the target and searching the target location which matches the characteristics in the follow-up frame based on the objectives and their environment. It is not only the main data source of target motion and scene analysis, but also of help for target detection and recognition. At present there are many algorithms on the target tracking (object tracking).Fundamental difference among the various algorithms is the choice of the characteristics of the target description and the type of search matching algorithm. In this paper, we introduce the particle filter algorithm.
The use of a combination of two techniques to achieve the improvement of the gait recognition rate is the main content and innovation of this article, the proposed method is:First, we use background subtraction method to extract the target from the background, and its Fourier descriptor features and then to use a particle filter algorithm to get a more accurate target location information in the process of movement, and on this basis, calculate Fourier descriptors to achieve superior gait recognition effect. It proves that the information extracted features with track is more accurate by the contrast of Fourier descriptor before and after filtering and the real target. This algorithm can improve the efficiency of target recognition of the human body.
引文
[1]魏怡,倪海峰.粒子滤波的技术及应用.中国水运.2009,09(3)：126-127.
[2]Shutler J, Nixon M, Harris C. Statistical gait recognition via temporal moments. In:Proc IEEE Southwest Symposium on Image Analysis and Interpretation, Austin, Texas,2000:291-295
[3]Hayfron-Acquah J, Nixon M, Carter J. Automatic gait recognition by symmetry analysis. In:Proc Int Con Audio-Video-Based Biometric Person Authentication,2001:272-277
[4]Ben-Abdelkader C, Cutler R, Davis L. Motion-based recognition of people in eigengait space. In:Proceedings of the Fifth IEEE International Conference on Automatic Face and Gesture Recognition,2002:254-259
[5]Collins RT, Cross P, Shi JB. Silhouette-based human identification from body shape and gait [A]. In:Proceedings of Fifth IEEE International Conference on Automatic Face and Gesture Recognition[C], Washington DC, USA,2002:351-356
[6]Foster J P, Nixon MS, Pruge-Bennet A. Automatic gait recognition using area based metrics. Pattern Recognition Letters,2003;24:2489
[7]Begg RK, Palaniswami M, Owen B. Support vector machines for automated gait classification. IEEE Transactions on Bio-medical Engineering,2005; 52 (5): 828
[8]Wang L, Tan TN, Hu WM, et al. Automatic gait recognition based on statistical shape analysis. IEEE Transaction on Image Processing,2003;12 (9):1120
[9]韩鸿哲,李彬,王志良等.基于傅立叶描述子的步态识别.计算机工程,200531(2) ：48
[10]田光见,赵荣椿.基于连续隐马尔可夫模型的步态识别.中国图像图形学报,2006；11(6)：867
[11]胡雪艳,恽晓平.步态分析在临床中的应用.中国康复理论与实践,2003,9 (11)：677
[12]M. Sanjeev Arulampalam, Simon Maskell, Neil Gordon, and Tim Clapp.A Tutorial on Particle Filters for Online Nonlinear/Non-Gaussian Bayesian Tracking. University of Electronic Science and Tech of China. Downloaded on November 19, 2008 at 21:47 from IEEE Xplore. Restrictions apply.
[13WU Y. Robust visual tracking by integrating multiple cues based on co-inference learning[J]. International Journal of Computer Vision,2004,58(1): 55-71.
[14]VIOLA P, JONES M J. Robust real-time face detection[J]. International Journal Of computer Vision,2004,52(2):137-154.
[15]Pan Pan, Student Member, IEEE, and Dan Schonfeld, Senior Member, IEEE, Dynamic Proposal Variance and Optimal Particle Allocation in Particle Filtering for Video Tracking, IEEE transactions on circuits and systems for video technology, 2008.09:1268-1279.
[16]Kass M, Witkin A, Terzopoulos D. Snakes:active contour models[J].International Journal of Computer Vision,1987,1(4):321-331
[17]王素玉,沈兰荪.智能视觉监控技术研究进展.中国图像图形学报.2007,12(9)：1505-1514.
[18]Ivins J, Porrill J. Active region models for segmenting textures and colours[J]. Image and Vision Computing,1995,13(5):431-438
[19]Hyunwook Park, Schopflin Todd, Yognin Kim. Active contour model with gradient directional information:directional snake [J]. IEEE Transactions on Circuits and Systems for Video Technology,2001,11 (2):252-256.
[20]Wojtaszek Daniel, Laganiere Robert. Tracking and recognition people in color using the earth mover's distance [A]. In:Proceedings of IEEE International Workshop on Hap tic Virtual Environments and Their Applications[C], Ottawa, Canada,2002: 91-96.
[21]Ramp rasad Polana, Randal Nelson. Low level recognition of human motion. Motion of Non2Rigid and Articulated Objects [A]. In:Proceedings of the 1994 IEEE Workshop on Motion of Non-Rigid and Articulated Objects[C], Texas, USA,1994: 77-82.
[22]Warren F Gardner, Daryl T Lawton. Interactive model-based vehicle tracking[J]. IEEE Transactions on Pattern Analysis and Intelligence,1996,18 (11): 1115-1121.
[23]Lou J ian-guang, Tan Tie-niu, HuWei-ming, et al.32-D model-based vehicle tracking[J]. IEEE Transactions on Image Processing,2005,14 (10):1561-1569
[24]Ralf Plankers, Pascal Fua. Tracking and modeling people in video sequences[J]. Computer Vision and Image Understanding,2001,81:285-302.
[25]鲍齐克,凌云旦.数字滤波和卡尔曼滤波[M].北京：科学出版社,1984.
[26]王亮,胡卫明,谭铁牛.基于步态的身份识别[J].北京：计算机学报,2003,26(3)：353-360.
[27]Hu Wei-ming, Tan Tie-niu, Steve Maybank. A survey on visual surveillance of object motion and behaviors[J]. IEEE Transactions on Systems, Man and Cybernetics2-Part C:Applications and Reviews,2004,34 (3):334-352.
[28]Stauffer Chris, GrimsonW E L. Adaptive background mixture models for real-time tracking[A]. In:Proceedings of IEEE Computer Society Conference on Computer Vision and Pattern Recognition Vol.2 [C],Colorado, USA,1999: 1063-1069.
[29]刘亚,艾海舟,徐光祐.一种基于背景模型的运动目标检测与跟踪算法[J].信息与控制,2002,31(4)：315～319.
[30]Karmann Klaus-Peter, Brandt Achim Von. Moving object recognition using an adaptive background memory[A]. In:CappelliniV. Time-Varying Image Processing and Moving Object Recognition [C],2,Elsevier, Amsterdam, The Netherlands,1990: 289～296.
[31]Kilger M. A shadow handler in a video-based real-time traffic monitoring system [A]. In:Proceedings of IEEE Workshop on Applications of Computer Vision[C], California, USA,1992:11-18.
[32]Haritaoglu Ismail, Harwood David, Davis Larry S. W4:who? when? where? what? a real time system for detecting and tracking people [A]. In:Proceedings of Third IEEE International Conference on Automatic Face and Gesture Recognition [C], Nara, Japan,1998:222～227.
[33]Jain R, Nagel H H. On the analysis of accumulative difference of picture from image sequences of real world scenes [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence,1979,PAM 1-1 (1):206-214.
[34]王东升,李在铭.空域视频场景监视中运动对象的实时检测与跟踪技术[J].信号处理,2005,21(2)：194-198.
[35]Horn Berthold K P, Schunck Brain G. Determining optical flow[J].Artificial Intelligence,1981,17 (1-3):185-203.
[36]Lucas B, Kanade T. An iterative image registration technique with an application to stereo vision [A]. In:Proceedings of DARPA Image Understanding Workshop [C], Pittsburgh, Pennsylvania, USA,1981:121-130.
[37]Nagel Hans Hellmut. Constrains for the estimation of displacement vector fields from image sequences [A]. In:Proceedings of International Joint Conference on Artificial Intelligence (IJCA 1283) [C], Karlsruhe, Germany,1983:945-951.
[38]Otte M, Nagel H H. Optical flow estimation:Advances and comparisons [A]. In:Eklundh J. Proceedings of European Conference on Computer Vision, volume 800 of LNCS2Series [C], Stockholm, Sweden,1994:51-60.
[39]Barron J L, Fleet D J, Beauchem in S S, et al. Performance of optical flow techniques [A]. In:Proceedings of IEEE Computer Society Conference on Computer Vision and Pattern Recognition [C], Illinois, USA,1992:236-242.
[40]Stephane Mallat, Wen Liang Hwang. Singularity detection and processing with wavelets [J]. IEEE Transactions on Information Theory,1992,38 (2):617-643.
[41]Jagdish Chandra Patra, Ferdinan Widjaja, et al. A fast neural network-based dection and tracking of dim moving targets in FLIR imagery[A]. Proceedings of International Joint Conference on Neural Networks[C], Montreal, Canada,2005: 3144-3149.
[42]Stilla U, Michaelsen E, Soergel U, et al. Perceptual grouping of regular structures for automatic detection of man2made objects [A].In:Proceedings of IEEE International Geoscience and Remote Sensing Symposium (IGARSS'03) [C] Toulouse, France,2003,6:3525-3527.
[43]A. Doucet, S.Godsill, Andrieu. On sequential Monte Carlo sampling methods for Bayesian filtering[J].Statist.Computer,2000,10:197-208
[44]S.A rulampalam and B.Ristic. Comparision of the particle filter with range parameterized and modified polar EKF's for angle-only tracking [J]. Signal and Data Processing of Small Targets, SPIE,2000,4048:288-299
[45]袁天鑫,最佳估计原理[M].北京：国防工业出版社,1980
[46]P. Del Moral. Measure valud processes and interacting particle systems. Application to nonlinear filtering problems. A nn. Appl. Probab.1998,8 (2)
[47]李庆扬,关治,白峰杉.数值计算原理[M].北京：清华大学出版社,2000
[48]A.Doucet, S.Godsill, Andrieu. On sequential Monte Carlo sampling methods for Bayesian filtering[J].St atist.Computer,2000,10:197-208
[49]Simon Godsill,Tim Clapp. Improvement strategies for Monte Carlo particle filters [D].Signal Processing Group, University of Cambridge
[50]Cunado D, Nixon M, Carter J. Using gait as a biometric, via phase-weighted magnitude spectra. In:Processings of International Conference on Audio-and Video-based Biometric Person Authentication, Crans-Montana, Switzerland,1997: 95-102
[51]Yoo Jang-Hee, Nixon M S, Harris C J. Extracting human gait signatures by body segment properties. Proceedings of IEEE Southwest Symposium on Image Analysis and Interpretation.2002,35-39
[52]Little J, Boyd J. Recognizing people by their gait:the shape of motion. Journal of Computer Vision Research,1998;1(2):2-32
[53]许文芳,吴清江.步态识别综述.福建电脑.2007(1)：19-20
[54]Zahn C T, Roskies R Z. Fourier Descriotors for Plane Closed Curves. IEEE Trans on Computers,1972,21(3):269-281
[55]杨路明,桂彦,曾庆冬.基于无标记点运动跟踪的步态参数提取[J].成都：计算机应用研究,2008,25(6)：1759-1764.
[56]Lu Xiaoqi, Huang Hongmei, Zhang Baohua. Recognition of human movement Based on Fourier Descriptors. Bioinformatics and Binmedical Engineering,2008, ICBBE 2008.The 2nd International Conference on Department of Communication Engineering, Engineering College of Armed Police Force..
[57]程琼,庄留杰,付波,基于傅里叶描述子和人工神经网络的步态识别,武汉：武汉理工大学学报,2008,30(1)：126-129.
[2]Shutler J, Nixon M, Harris C. Statistical gait recognition via temporal moments. In:Proc IEEE Southwest Symposium on Image Analysis and Interpretation, Austin, Texas,2000:291-295
[3]Hayfron-Acquah J, Nixon M, Carter J. Automatic gait recognition by symmetry analysis. In:Proc Int Con Audio-Video-Based Biometric Person Authentication,2001:272-277
[4]Ben-Abdelkader C, Cutler R, Davis L. Motion-based recognition of people in eigengait space. In:Proceedings of the Fifth IEEE International Conference on Automatic Face and Gesture Recognition,2002:254-259
[5]Collins RT, Cross P, Shi JB. Silhouette-based human identification from body shape and gait [A]. In:Proceedings of Fifth IEEE International Conference on Automatic Face and Gesture Recognition[C], Washington DC, USA,2002:351-356
[6]Foster J P, Nixon MS, Pruge-Bennet A. Automatic gait recognition using area based metrics. Pattern Recognition Letters,2003;24:2489
[7]Begg RK, Palaniswami M, Owen B. Support vector machines for automated gait classification. IEEE Transactions on Bio-medical Engineering,2005; 52 (5): 828
[8]Wang L, Tan TN, Hu WM, et al. Automatic gait recognition based on statistical shape analysis. IEEE Transaction on Image Processing,2003;12 (9):1120
[9]韩鸿哲,李彬,王志良等.基于傅立叶描述子的步态识别.计算机工程,200531(2) ：48
[10]田光见,赵荣椿.基于连续隐马尔可夫模型的步态识别.中国图像图形学报,2006；11(6)：867
[11]胡雪艳,恽晓平.步态分析在临床中的应用.中国康复理论与实践,2003,9 (11)：677
[12]M. Sanjeev Arulampalam, Simon Maskell, Neil Gordon, and Tim Clapp.A Tutorial on Particle Filters for Online Nonlinear/Non-Gaussian Bayesian Tracking. University of Electronic Science and Tech of China. Downloaded on November 19, 2008 at 21:47 from IEEE Xplore. Restrictions apply.
[13WU Y. Robust visual tracking by integrating multiple cues based on co-inference learning[J]. International Journal of Computer Vision,2004,58(1): 55-71.
[14]VIOLA P, JONES M J. Robust real-time face detection[J]. International Journal Of computer Vision,2004,52(2):137-154.
[15]Pan Pan, Student Member, IEEE, and Dan Schonfeld, Senior Member, IEEE, Dynamic Proposal Variance and Optimal Particle Allocation in Particle Filtering for Video Tracking, IEEE transactions on circuits and systems for video technology, 2008.09:1268-1279.
[16]Kass M, Witkin A, Terzopoulos D. Snakes:active contour models[J].International Journal of Computer Vision,1987,1(4):321-331
[17]王素玉,沈兰荪.智能视觉监控技术研究进展.中国图像图形学报.2007,12(9)：1505-1514.
[18]Ivins J, Porrill J. Active region models for segmenting textures and colours[J]. Image and Vision Computing,1995,13(5):431-438
[19]Hyunwook Park, Schopflin Todd, Yognin Kim. Active contour model with gradient directional information:directional snake [J]. IEEE Transactions on Circuits and Systems for Video Technology,2001,11 (2):252-256.
[20]Wojtaszek Daniel, Laganiere Robert. Tracking and recognition people in color using the earth mover's distance [A]. In:Proceedings of IEEE International Workshop on Hap tic Virtual Environments and Their Applications[C], Ottawa, Canada,2002: 91-96.
[21]Ramp rasad Polana, Randal Nelson. Low level recognition of human motion. Motion of Non2Rigid and Articulated Objects [A]. In:Proceedings of the 1994 IEEE Workshop on Motion of Non-Rigid and Articulated Objects[C], Texas, USA,1994: 77-82.
[22]Warren F Gardner, Daryl T Lawton. Interactive model-based vehicle tracking[J]. IEEE Transactions on Pattern Analysis and Intelligence,1996,18 (11): 1115-1121.
[23]Lou J ian-guang, Tan Tie-niu, HuWei-ming, et al.32-D model-based vehicle tracking[J]. IEEE Transactions on Image Processing,2005,14 (10):1561-1569
[24]Ralf Plankers, Pascal Fua. Tracking and modeling people in video sequences[J]. Computer Vision and Image Understanding,2001,81:285-302.
[25]鲍齐克,凌云旦.数字滤波和卡尔曼滤波[M].北京：科学出版社,1984.
[26]王亮,胡卫明,谭铁牛.基于步态的身份识别[J].北京：计算机学报,2003,26(3)：353-360.
[27]Hu Wei-ming, Tan Tie-niu, Steve Maybank. A survey on visual surveillance of object motion and behaviors[J]. IEEE Transactions on Systems, Man and Cybernetics2-Part C:Applications and Reviews,2004,34 (3):334-352.
[28]Stauffer Chris, GrimsonW E L. Adaptive background mixture models for real-time tracking[A]. In:Proceedings of IEEE Computer Society Conference on Computer Vision and Pattern Recognition Vol.2 [C],Colorado, USA,1999: 1063-1069.
[29]刘亚,艾海舟,徐光祐.一种基于背景模型的运动目标检测与跟踪算法[J].信息与控制,2002,31(4)：315～319.
[30]Karmann Klaus-Peter, Brandt Achim Von. Moving object recognition using an adaptive background memory[A]. In:CappelliniV. Time-Varying Image Processing and Moving Object Recognition [C],2,Elsevier, Amsterdam, The Netherlands,1990: 289～296.
[31]Kilger M. A shadow handler in a video-based real-time traffic monitoring system [A]. In:Proceedings of IEEE Workshop on Applications of Computer Vision[C], California, USA,1992:11-18.
[32]Haritaoglu Ismail, Harwood David, Davis Larry S. W4:who? when? where? what? a real time system for detecting and tracking people [A]. In:Proceedings of Third IEEE International Conference on Automatic Face and Gesture Recognition [C], Nara, Japan,1998:222～227.
[33]Jain R, Nagel H H. On the analysis of accumulative difference of picture from image sequences of real world scenes [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence,1979,PAM 1-1 (1):206-214.
[34]王东升,李在铭.空域视频场景监视中运动对象的实时检测与跟踪技术[J].信号处理,2005,21(2)：194-198.
[35]Horn Berthold K P, Schunck Brain G. Determining optical flow[J].Artificial Intelligence,1981,17 (1-3):185-203.
[36]Lucas B, Kanade T. An iterative image registration technique with an application to stereo vision [A]. In:Proceedings of DARPA Image Understanding Workshop [C], Pittsburgh, Pennsylvania, USA,1981:121-130.
[37]Nagel Hans Hellmut. Constrains for the estimation of displacement vector fields from image sequences [A]. In:Proceedings of International Joint Conference on Artificial Intelligence (IJCA 1283) [C], Karlsruhe, Germany,1983:945-951.
[38]Otte M, Nagel H H. Optical flow estimation:Advances and comparisons [A]. In:Eklundh J. Proceedings of European Conference on Computer Vision, volume 800 of LNCS2Series [C], Stockholm, Sweden,1994:51-60.
[39]Barron J L, Fleet D J, Beauchem in S S, et al. Performance of optical flow techniques [A]. In:Proceedings of IEEE Computer Society Conference on Computer Vision and Pattern Recognition [C], Illinois, USA,1992:236-242.
[40]Stephane Mallat, Wen Liang Hwang. Singularity detection and processing with wavelets [J]. IEEE Transactions on Information Theory,1992,38 (2):617-643.
[41]Jagdish Chandra Patra, Ferdinan Widjaja, et al. A fast neural network-based dection and tracking of dim moving targets in FLIR imagery[A]. Proceedings of International Joint Conference on Neural Networks[C], Montreal, Canada,2005: 3144-3149.
[42]Stilla U, Michaelsen E, Soergel U, et al. Perceptual grouping of regular structures for automatic detection of man2made objects [A].In:Proceedings of IEEE International Geoscience and Remote Sensing Symposium (IGARSS'03) [C] Toulouse, France,2003,6:3525-3527.
[43]A. Doucet, S.Godsill, Andrieu. On sequential Monte Carlo sampling methods for Bayesian filtering[J].Statist.Computer,2000,10:197-208
[44]S.A rulampalam and B.Ristic. Comparision of the particle filter with range parameterized and modified polar EKF's for angle-only tracking [J]. Signal and Data Processing of Small Targets, SPIE,2000,4048:288-299
[45]袁天鑫,最佳估计原理[M].北京：国防工业出版社,1980
[46]P. Del Moral. Measure valud processes and interacting particle systems. Application to nonlinear filtering problems. A nn. Appl. Probab.1998,8 (2)
[47]李庆扬,关治,白峰杉.数值计算原理[M].北京：清华大学出版社,2000
[48]A.Doucet, S.Godsill, Andrieu. On sequential Monte Carlo sampling methods for Bayesian filtering[J].St atist.Computer,2000,10:197-208
[49]Simon Godsill,Tim Clapp. Improvement strategies for Monte Carlo particle filters [D].Signal Processing Group, University of Cambridge
[50]Cunado D, Nixon M, Carter J. Using gait as a biometric, via phase-weighted magnitude spectra. In:Processings of International Conference on Audio-and Video-based Biometric Person Authentication, Crans-Montana, Switzerland,1997: 95-102
[51]Yoo Jang-Hee, Nixon M S, Harris C J. Extracting human gait signatures by body segment properties. Proceedings of IEEE Southwest Symposium on Image Analysis and Interpretation.2002,35-39
[52]Little J, Boyd J. Recognizing people by their gait:the shape of motion. Journal of Computer Vision Research,1998;1(2):2-32
[53]许文芳,吴清江.步态识别综述.福建电脑.2007(1)：19-20
[54]Zahn C T, Roskies R Z. Fourier Descriotors for Plane Closed Curves. IEEE Trans on Computers,1972,21(3):269-281
[55]杨路明,桂彦,曾庆冬.基于无标记点运动跟踪的步态参数提取[J].成都：计算机应用研究,2008,25(6)：1759-1764.
[56]Lu Xiaoqi, Huang Hongmei, Zhang Baohua. Recognition of human movement Based on Fourier Descriptors. Bioinformatics and Binmedical Engineering,2008, ICBBE 2008.The 2nd International Conference on Department of Communication Engineering, Engineering College of Armed Police Force..
[57]程琼,庄留杰,付波,基于傅里叶描述子和人工神经网络的步态识别,武汉：武汉理工大学学报,2008,30(1)：126-129.