基于Hi3510的视频服务器软硬件平台开发与移动侦测算法实现
|
摘要
随着信息化时代的全面到来,视频服务器技术得到了长足的发展,应用领域也越来越广泛,例如监控领域、网络教育、网络点播、视频会议、远程医疗会诊等,随着视频服务器技术向着前端一体化、视频数字化、控制网络化、系统集成化的方向发展,其应用价值将会得到更加充分的体现。
本文以吉林省科技厅发展计划项目为背景,主要内容是完成新一代高性能网络视频服务器的产品开发。本文首先深入研究了多媒体处理芯片Hi3510的性能特征,进行了硬件系统电路的设计,包括原理图的设计和PCB的绘制;其次在硬件电路基础上进行了嵌入式Linux系统开发环境的搭建,并在该环境下进行了设备驱动程序以及服务器端的应用程序设计,采用C/S软件系统架构进行音视频实时流媒体传输以及服务器与客户端之间的网络交互,实现了项目预计的所有功能,即服务器端可以采集四路音视频信号进行H.264视频编码和G.726音频编码,并通过音视频复合通道进行网络流媒体传输,支持多客户端同时访问、多种视频格式(1D1、4CIF等)网络点播、音频播放、网络云台控制、远程告警输入/输出监测、服务器与客户端的信息交互等网络监控功能;最后在此基础上研究了移动侦测算法,并通对驱动和用户程序的改进,实现了移动侦测功能,为新一代智能监控系统的搭建提供了有效的解决方案。
With the comprehensive arrival of the information age, the core technology of the monitor system develops from the traditional analog system towards a new generation of digital and network-based system, the monitoring requirements have become more sophisticated, front-end integration, digital video, monitoring network, and system integration of video surveillance system is recognized as the direction of development, audio and video server technology is a cross subject collecting the audio and video codecs, network communications, streaming media processing, artificial intelligence as a whole, the subject is from Science and Technology Department of Jilin Province, the main contents of the subject is to complete the the development of high-performance audio and video server products, the products use Hi3510 multimedia processing chip provided by Hisilicon as the core processor, use embedded Linux operating system as the system architecture,which can be mixed on the four-way streaming audio and video transmission, use H.264 codec for video, G.726 codec foe voice, the overall use C / S network management model that can accept multi-user online access at the same time, through the RTSP, RTP / RTCP protocol achieve a network of multi-functional on-demand streaming media, through the TCP protocol server and client interact under network monitoring functions, and under such platform the mobile detection function is achieved, which provides solutions for the new generation of intelligent monitoring system.
The task of author in the subject is the main server-side hardware and software platform design and the realization of mobile detection algorithm.Firstly study performance characteristics of the multimedia processing chip Hi3510, Hi3510 has the ARM + DSP dual-core SOC architecture, which integrate an ARM926EJ-S and associated peripherals, as well as a high-performance DSP co-processor. The ARM926EJ-S processor works as a master chip,in which a variety of user interface, application software, operating system and protocol-processing software are running. DSP system is responsible for the audio codec processes, co-processor cooperate with video codec module to complete H.261/H.263/H.264 codec.
Based on the processing core Hi3510 design the overall of hardware circuit, through the tools of the Protel 99SE complete the schematic circuit design and the realization of 6 layers PCB, the circuit schematic design includes package design, electrical properties of pin settings, connections between components, network configuration label, ERC (Electrical rule checking), netlist generation and so on, then followed the device PCB package design,and build relationship between schematics components and PCB package, the next step is importing netlist into PCB design moudle, followed with the circuit boards planning, placement and routing, DRC (design rule checking), the division of ground and power, and ultimately video and audio server hardware systems come ture.
On the hardware platform the embedded Linux system is built, the composition of the system are: bootloader, Linux kernel ,initialization process, the hardware drivers and the application. Bootloader is similar to the computer's BIOS to boot the system, it will drive CPU, network chips and other basic hardware devices, Linux kernel will be loaded into memory, and make the kernel initialization process control the system, generaly ARM system selected U-boot as bootloader, here use the improved version Hiboot for the hardware system; kernel is the core component of the system, its function include memory management, process management, and timer interrupt management, module management, interface file system, device drivers, inter-process communication, network management, etc.; initialization process, the hardware driver, the application are the upper parts, which depend on kernel scheduling, but have their own separate mechanism to deal with, it is the system's main components to achieve the functions, they are the root file system's main components.
Linux root file system has many formats, cramfs and jffs2 are commonly used, but they all belong to the binary image file, which need to be written to the FLASH to implement, and during the development stage NFS(Network File System) is a more effective way, it is visited by the target machine by the means of a network shared file system, which is seemed as the root file system for its application to the implementation of this development environment, known as cross-compiler environment, in such an environment can be easily the conduct of drivers and applications debugging. Were carried out in the design of the TW2834, gpio, such as character device driver design, the realization of a four-way All the way to deal with a variety of video output of video processing mode and the alarm signal input and output, and drive through the RS485 PTZ pelco agreement decoder, the realization of computer-controlled PTZ camera rotation and adjustment of program design in which the user using C / S of the network model, using multi-threaded structure of modular design, each functional module, respectively, to achieve the audio and video data collection, coding, network transmission and user control functions, data queue between threads, as well as the amount mutually exclusive mechanisms of communication and synchronization. Which, through RTSP, RTP / RTCP to achieve a real-time streaming media transmission network through TCP / IP protocol and the client server network interaction, remote control in order to achieve a streaming media format NTSC and PAL under the 1D1, 4CIF video format, such as on-demand, as well as alarm input and output device, PTZ control functions such as network monitoring.
In addition, the article carried out extensive research for motion detection algorithm, to consider the feasibility of the realization of projects , the difference image algorithm is studied deeply and achieved in the user program. The algorithm achievement is based on TW2834 for its support of motion detection ,the algorithm based on difference image processing method: Detect the differences of the brightness level between the current frame and reference frame, to detect moving objects in images, server-side through the network thread accept and implement the control command of client-side to detect the motion, the client can the set up motion detection region for each video channel, the motion detection alarm information output to the user interface, while the information is also saved to documents in the system for the future calls.
In this paper ,the products of the project has passed the inspection, having outstanding performance, to achieve the targets, which has achieved the functions required by the four-way audio and video server, the product can be playing an important role in the new generation of network monitoring systems.
本文以吉林省科技厅发展计划项目为背景,主要内容是完成新一代高性能网络视频服务器的产品开发。本文首先深入研究了多媒体处理芯片Hi3510的性能特征,进行了硬件系统电路的设计,包括原理图的设计和PCB的绘制;其次在硬件电路基础上进行了嵌入式Linux系统开发环境的搭建,并在该环境下进行了设备驱动程序以及服务器端的应用程序设计,采用C/S软件系统架构进行音视频实时流媒体传输以及服务器与客户端之间的网络交互,实现了项目预计的所有功能,即服务器端可以采集四路音视频信号进行H.264视频编码和G.726音频编码,并通过音视频复合通道进行网络流媒体传输,支持多客户端同时访问、多种视频格式(1D1、4CIF等)网络点播、音频播放、网络云台控制、远程告警输入/输出监测、服务器与客户端的信息交互等网络监控功能;最后在此基础上研究了移动侦测算法,并通对驱动和用户程序的改进,实现了移动侦测功能,为新一代智能监控系统的搭建提供了有效的解决方案。
With the comprehensive arrival of the information age, the core technology of the monitor system develops from the traditional analog system towards a new generation of digital and network-based system, the monitoring requirements have become more sophisticated, front-end integration, digital video, monitoring network, and system integration of video surveillance system is recognized as the direction of development, audio and video server technology is a cross subject collecting the audio and video codecs, network communications, streaming media processing, artificial intelligence as a whole, the subject is from Science and Technology Department of Jilin Province, the main contents of the subject is to complete the the development of high-performance audio and video server products, the products use Hi3510 multimedia processing chip provided by Hisilicon as the core processor, use embedded Linux operating system as the system architecture,which can be mixed on the four-way streaming audio and video transmission, use H.264 codec for video, G.726 codec foe voice, the overall use C / S network management model that can accept multi-user online access at the same time, through the RTSP, RTP / RTCP protocol achieve a network of multi-functional on-demand streaming media, through the TCP protocol server and client interact under network monitoring functions, and under such platform the mobile detection function is achieved, which provides solutions for the new generation of intelligent monitoring system.
The task of author in the subject is the main server-side hardware and software platform design and the realization of mobile detection algorithm.Firstly study performance characteristics of the multimedia processing chip Hi3510, Hi3510 has the ARM + DSP dual-core SOC architecture, which integrate an ARM926EJ-S and associated peripherals, as well as a high-performance DSP co-processor. The ARM926EJ-S processor works as a master chip,in which a variety of user interface, application software, operating system and protocol-processing software are running. DSP system is responsible for the audio codec processes, co-processor cooperate with video codec module to complete H.261/H.263/H.264 codec.
Based on the processing core Hi3510 design the overall of hardware circuit, through the tools of the Protel 99SE complete the schematic circuit design and the realization of 6 layers PCB, the circuit schematic design includes package design, electrical properties of pin settings, connections between components, network configuration label, ERC (Electrical rule checking), netlist generation and so on, then followed the device PCB package design,and build relationship between schematics components and PCB package, the next step is importing netlist into PCB design moudle, followed with the circuit boards planning, placement and routing, DRC (design rule checking), the division of ground and power, and ultimately video and audio server hardware systems come ture.
On the hardware platform the embedded Linux system is built, the composition of the system are: bootloader, Linux kernel ,initialization process, the hardware drivers and the application. Bootloader is similar to the computer's BIOS to boot the system, it will drive CPU, network chips and other basic hardware devices, Linux kernel will be loaded into memory, and make the kernel initialization process control the system, generaly ARM system selected U-boot as bootloader, here use the improved version Hiboot for the hardware system; kernel is the core component of the system, its function include memory management, process management, and timer interrupt management, module management, interface file system, device drivers, inter-process communication, network management, etc.; initialization process, the hardware driver, the application are the upper parts, which depend on kernel scheduling, but have their own separate mechanism to deal with, it is the system's main components to achieve the functions, they are the root file system's main components.
Linux root file system has many formats, cramfs and jffs2 are commonly used, but they all belong to the binary image file, which need to be written to the FLASH to implement, and during the development stage NFS(Network File System) is a more effective way, it is visited by the target machine by the means of a network shared file system, which is seemed as the root file system for its application to the implementation of this development environment, known as cross-compiler environment, in such an environment can be easily the conduct of drivers and applications debugging. Were carried out in the design of the TW2834, gpio, such as character device driver design, the realization of a four-way All the way to deal with a variety of video output of video processing mode and the alarm signal input and output, and drive through the RS485 PTZ pelco agreement decoder, the realization of computer-controlled PTZ camera rotation and adjustment of program design in which the user using C / S of the network model, using multi-threaded structure of modular design, each functional module, respectively, to achieve the audio and video data collection, coding, network transmission and user control functions, data queue between threads, as well as the amount mutually exclusive mechanisms of communication and synchronization. Which, through RTSP, RTP / RTCP to achieve a real-time streaming media transmission network through TCP / IP protocol and the client server network interaction, remote control in order to achieve a streaming media format NTSC and PAL under the 1D1, 4CIF video format, such as on-demand, as well as alarm input and output device, PTZ control functions such as network monitoring.
In addition, the article carried out extensive research for motion detection algorithm, to consider the feasibility of the realization of projects , the difference image algorithm is studied deeply and achieved in the user program. The algorithm achievement is based on TW2834 for its support of motion detection ,the algorithm based on difference image processing method: Detect the differences of the brightness level between the current frame and reference frame, to detect moving objects in images, server-side through the network thread accept and implement the control command of client-side to detect the motion, the client can the set up motion detection region for each video channel, the motion detection alarm information output to the user interface, while the information is also saved to documents in the system for the future calls.
In this paper ,the products of the project has passed the inspection, having outstanding performance, to achieve the targets, which has achieved the functions required by the four-way audio and video server, the product can be playing an important role in the new generation of network monitoring systems.
引文
[1]苏州科达科技有限公司,科达企业级视频监控系统解决方案,数字社区&智能家居, 2007(3).
[2]赵红平,网络视频面临的新挑战,安讯士网络通讯有限公司, 2009.
[3]苏州科达科技有限公司,运营级视频监控的视频编解码与流媒体分发技术研究和应用,数字社区&智能家居, 2007(3).
[4]谭晓,明明,嵌入式Linux平台的网络视频监控系统,电子技术, 2006.
[5] ISO/IEC CD 13818, Generic Coding of Moving Pictures and Associated Audio Information, Nov., 1994.
[6] ISO/IEC CD 14496, Final Draft International Standard, Information Technology-Generic Coding of Audio-Visual Objects, Dec. 1998.
[7] R. Talluri, Error-Resilient Video Coding in the ISO MPEG-4 Standard, IEEE communications Magazine, Vol. 36 No. 6, June 1998.
[8] ITU-T Recommecdation H.263, Video Coding for Low Bitrate Communication, 1995.
[9] ITU-T Recommecdation H.264, Advanced Video Coding for Generic Audio-Visual Services, 2003.
[10] CCITT SGXV Recommendation H.261, Video Codec for Audio-Visual Services at p*64 kbit/s, Aug. 1990.
[11] CCITT Recommendation G.711, Pulse Code Modulation (PCM) of Voice Frequencies, Aug. 1972.
[12] CCITT Recommendation G.721, 32 kbit/s Adaptive Differential Pulse Code Modulation (ADPCM), 1984.
[13] Lee Cheng-Chieh, An Enhanced ADPCM Coder for Voice Over Packet Networks [J], International Journal of Speech Technology, 1999, 2(4), 343 -357.
[14]李善平,刘文峰,王焕龙, Linux与嵌入式系统,清华大学出版社, 2006.
[15] Hi3510 HiBoot移植应用Application Notes,深圳市海思半导体有限公司, 2006.
[16] Hi3510 HiBoot与标准U-boot差异说明,深圳市海思半导体有限公司, 2006.
[17]毛德操胡希明, Linux内核源代码情景分析[M],浙江:浙江大学出版社, 2001: 1-21.
[18] Jonathan Corbet等著, Linux设备驱动程序(第三版)[M],北京:中国电力出版社, 2007:9-19.
[19]郑家玲,张云峰,孙荷琨,嵌入式linux内核下串行驱动程序的实现,西安电子科技大学信息科学研究所, 2002.
[20] Adolfo Rodriguez, John Gatrell, TCP/IP Tutorial and Technical Overview 7th Edition, Pearson Education Press, 2002.
[21] T. Socolofsky, C.Kale, TCP/IP tutorial RFC1180, 1991.
[22] UDP: User Datagram Protocol, http://www.javvin.com/protocolUDP.html.
[23] RTP: A Transport Protocol for Real-Time Applications RFC 1889, IETF, 1999.
[24] H.Schulzrinne, S.Casner, R.Frederick, A Transport Protocol for Real-Time Applications, RFC3550, 2003.
[25] Pelco-D Protocol Manual, Pelco Company, 1999.
[26] Dagan E, Mano O, Stein G P, et al, Forward collision warning with a single camera[C]//IEEE Intelligent Vehicles Symposium, June 2004: 37-42.
[27] Feraric J P, Onken R Daisy, A driver assisting system which adapts to the driver[C]//IFAC Man- Machine Systems Symposium, 1995: 523-528.
[28] Iteris’lane departure warning system now available on Mercedes trucks in Europe, The Source for Intelligent Vehicle News, Ivsource, 2000.
[29] Hi3510V100通信媒体处理器芯片用户指南,深圳市海思半导体有限公司, 2006.
[30] Spansion Corporation, S29GL-A MirrorBit? Flash Family Datasheet, 2004.
[31] Techwell, Inc. TW2834 4 Channel Video QUAD/MUX Controller for Security Applications, 2005: 14-28.
[32] Techwell, Inc.TW2834 4 Channel Video QUAD/MUX Controller for Security Applications, 2005: 237-239.
[33]高鹏,安涛,冠怀成,电路设计与制版Protel 99提高与应用[M],北京:人民邮电出版社, 2000: 1-27.
[34]赵晶,电路设计与制版Protel 99高级应用[M],北京:人民邮电出版社, 2001: 88-113.
[35] Hi3510 Linux Development Environment User Guide, Hisilicon Semiconductor Proprietary, 2006: 11-15.
[36] Hi3510 Linux开发环境用户指南,深圳市海思半导体有限公司, 2006.
[37] Karim Yaghmour1 Building embedded Linux system [M], O’Reilly & Associates, Inc12003.
[38]杜俊俐,王东云,基于C/S与B/S混合架构的远程监控系统研究[J],中原工学院学报, 2003, 14(1): 1-3.
[39] Hi3510媒体处理软件开发指南,深圳市海思半导体有限公司, 2006.
[40]王栓等,基于差分图像的多运动目标的检测与跟踪[J],中国图像图形学报, 1999, 4(A).
[41] Alan J Lipton, H Fujiyoshi, Raju S Patil, Moving target classification and tracking from real-time video[J], IEEE Transactions on Workshop Application of Computer Vision, 1998: 8-14.
[42] Alireza Behrad, Ali Shahrokni, Seyed Ahmad Motamedi, A robust vision-based moving target detection and tracking system[EB/OL], http://ligwww.epfl.ch/~ali.
[43] Jaewon Shin, Initialization of visual object tracker using frame absolute difference[EB/OL], http:∥ise.stanford.edu/class/ee392j/projects/shin-report.pdf.
[44] Paul L Rosin , Tim Ellis, Image difference threshold strategies and shadow detection[EB/OL], http:∥www.cs.cf.ac.uk/User/Paul.Rosin/resources/papers/shadows.pdf.
[45] Jain R, Nagel H, On the analysis of accumulative difference of picture from image sequences of real world scenes[J], IEEE Transactions PAMI, 1979: 206-214.
[46]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).
[47]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).
[48] Shoichi Araki, Takashi Matsuoaka, Naokazu Yokoya, et al, Realtime tracking of multiple moving object contours in a moving camera image sequence[J], IEICE Trans. Info. & Syst., 2000, E832D(7).
[49] Yasushi Mae, Yoshiaki Shirai, Jun Miura, et al, Object tracking in cluttered background based on optical flow and edges[EB/OL], http://www.cv.mech.eng.osakau.ac.jp/research/tracking-group/mae/flow- edge/ICPR96/ICPR96.html.
[50] Yasushi Mae, Shinya Yamamoto, Yoshiaki Shirai, et al, Opticalflow based realtime object tracking by active vision System [EB/OL], http://www.cv. mech. eng. osakau.ac.jp/research/tracking-group/mae/flow/J FCM94/J FCM94. html.
[51] Ryuzo Okada, Yoshiaki Shirai, Jun Miura, Object tracking based on optical flow and depth [EB/OL], http://www.cv.mech.eng.osakau.ac.jp/research/tracking-group/okada/flow-disp-tracking/flow-disp-tracking.html.
[52] Ryuzo Okada, Shinya Yamamoto, Yasushi Mae, Realtime person tracking system [EB/OL], http:∥www.cv.mech.eng.osakau.ac.jp/research/tracking-group/okada/person-tracking/paper/TIcontest.ps.gz.
[53]伏思华,张小虎,基于序列图像的运动目标实时检测方法[J],国防科学技术大学, 2004.
[2]赵红平,网络视频面临的新挑战,安讯士网络通讯有限公司, 2009.
[3]苏州科达科技有限公司,运营级视频监控的视频编解码与流媒体分发技术研究和应用,数字社区&智能家居, 2007(3).
[4]谭晓,明明,嵌入式Linux平台的网络视频监控系统,电子技术, 2006.
[5] ISO/IEC CD 13818, Generic Coding of Moving Pictures and Associated Audio Information, Nov., 1994.
[6] ISO/IEC CD 14496, Final Draft International Standard, Information Technology-Generic Coding of Audio-Visual Objects, Dec. 1998.
[7] R. Talluri, Error-Resilient Video Coding in the ISO MPEG-4 Standard, IEEE communications Magazine, Vol. 36 No. 6, June 1998.
[8] ITU-T Recommecdation H.263, Video Coding for Low Bitrate Communication, 1995.
[9] ITU-T Recommecdation H.264, Advanced Video Coding for Generic Audio-Visual Services, 2003.
[10] CCITT SGXV Recommendation H.261, Video Codec for Audio-Visual Services at p*64 kbit/s, Aug. 1990.
[11] CCITT Recommendation G.711, Pulse Code Modulation (PCM) of Voice Frequencies, Aug. 1972.
[12] CCITT Recommendation G.721, 32 kbit/s Adaptive Differential Pulse Code Modulation (ADPCM), 1984.
[13] Lee Cheng-Chieh, An Enhanced ADPCM Coder for Voice Over Packet Networks [J], International Journal of Speech Technology, 1999, 2(4), 343 -357.
[14]李善平,刘文峰,王焕龙, Linux与嵌入式系统,清华大学出版社, 2006.
[15] Hi3510 HiBoot移植应用Application Notes,深圳市海思半导体有限公司, 2006.
[16] Hi3510 HiBoot与标准U-boot差异说明,深圳市海思半导体有限公司, 2006.
[17]毛德操胡希明, Linux内核源代码情景分析[M],浙江:浙江大学出版社, 2001: 1-21.
[18] Jonathan Corbet等著, Linux设备驱动程序(第三版)[M],北京:中国电力出版社, 2007:9-19.
[19]郑家玲,张云峰,孙荷琨,嵌入式linux内核下串行驱动程序的实现,西安电子科技大学信息科学研究所, 2002.
[20] Adolfo Rodriguez, John Gatrell, TCP/IP Tutorial and Technical Overview 7th Edition, Pearson Education Press, 2002.
[21] T. Socolofsky, C.Kale, TCP/IP tutorial RFC1180, 1991.
[22] UDP: User Datagram Protocol, http://www.javvin.com/protocolUDP.html.
[23] RTP: A Transport Protocol for Real-Time Applications RFC 1889, IETF, 1999.
[24] H.Schulzrinne, S.Casner, R.Frederick, A Transport Protocol for Real-Time Applications, RFC3550, 2003.
[25] Pelco-D Protocol Manual, Pelco Company, 1999.
[26] Dagan E, Mano O, Stein G P, et al, Forward collision warning with a single camera[C]//IEEE Intelligent Vehicles Symposium, June 2004: 37-42.
[27] Feraric J P, Onken R Daisy, A driver assisting system which adapts to the driver[C]//IFAC Man- Machine Systems Symposium, 1995: 523-528.
[28] Iteris’lane departure warning system now available on Mercedes trucks in Europe, The Source for Intelligent Vehicle News, Ivsource, 2000.
[29] Hi3510V100通信媒体处理器芯片用户指南,深圳市海思半导体有限公司, 2006.
[30] Spansion Corporation, S29GL-A MirrorBit? Flash Family Datasheet, 2004.
[31] Techwell, Inc. TW2834 4 Channel Video QUAD/MUX Controller for Security Applications, 2005: 14-28.
[32] Techwell, Inc.TW2834 4 Channel Video QUAD/MUX Controller for Security Applications, 2005: 237-239.
[33]高鹏,安涛,冠怀成,电路设计与制版Protel 99提高与应用[M],北京:人民邮电出版社, 2000: 1-27.
[34]赵晶,电路设计与制版Protel 99高级应用[M],北京:人民邮电出版社, 2001: 88-113.
[35] Hi3510 Linux Development Environment User Guide, Hisilicon Semiconductor Proprietary, 2006: 11-15.
[36] Hi3510 Linux开发环境用户指南,深圳市海思半导体有限公司, 2006.
[37] Karim Yaghmour1 Building embedded Linux system [M], O’Reilly & Associates, Inc12003.
[38]杜俊俐,王东云,基于C/S与B/S混合架构的远程监控系统研究[J],中原工学院学报, 2003, 14(1): 1-3.
[39] Hi3510媒体处理软件开发指南,深圳市海思半导体有限公司, 2006.
[40]王栓等,基于差分图像的多运动目标的检测与跟踪[J],中国图像图形学报, 1999, 4(A).
[41] Alan J Lipton, H Fujiyoshi, Raju S Patil, Moving target classification and tracking from real-time video[J], IEEE Transactions on Workshop Application of Computer Vision, 1998: 8-14.
[42] Alireza Behrad, Ali Shahrokni, Seyed Ahmad Motamedi, A robust vision-based moving target detection and tracking system[EB/OL], http://ligwww.epfl.ch/~ali.
[43] Jaewon Shin, Initialization of visual object tracker using frame absolute difference[EB/OL], http:∥ise.stanford.edu/class/ee392j/projects/shin-report.pdf.
[44] Paul L Rosin , Tim Ellis, Image difference threshold strategies and shadow detection[EB/OL], http:∥www.cs.cf.ac.uk/User/Paul.Rosin/resources/papers/shadows.pdf.
[45] Jain R, Nagel H, On the analysis of accumulative difference of picture from image sequences of real world scenes[J], IEEE Transactions PAMI, 1979: 206-214.
[46]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).
[47]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).
[48] Shoichi Araki, Takashi Matsuoaka, Naokazu Yokoya, et al, Realtime tracking of multiple moving object contours in a moving camera image sequence[J], IEICE Trans. Info. & Syst., 2000, E832D(7).
[49] Yasushi Mae, Yoshiaki Shirai, Jun Miura, et al, Object tracking in cluttered background based on optical flow and edges[EB/OL], http://www.cv.mech.eng.osakau.ac.jp/research/tracking-group/mae/flow- edge/ICPR96/ICPR96.html.
[50] Yasushi Mae, Shinya Yamamoto, Yoshiaki Shirai, et al, Opticalflow based realtime object tracking by active vision System [EB/OL], http://www.cv. mech. eng. osakau.ac.jp/research/tracking-group/mae/flow/J FCM94/J FCM94. html.
[51] Ryuzo Okada, Yoshiaki Shirai, Jun Miura, Object tracking based on optical flow and depth [EB/OL], http://www.cv.mech.eng.osakau.ac.jp/research/tracking-group/okada/flow-disp-tracking/flow-disp-tracking.html.
[52] Ryuzo Okada, Shinya Yamamoto, Yasushi Mae, Realtime person tracking system [EB/OL], http:∥www.cv.mech.eng.osakau.ac.jp/research/tracking-group/okada/person-tracking/paper/TIcontest.ps.gz.
[53]伏思华,张小虎,基于序列图像的运动目标实时检测方法[J],国防科学技术大学, 2004.