基于NIOSII的局步运动目标提取及远程监控系统的设计与实现
|
摘要
近年来随着多媒体技术的迅速发展和计算机性能的不断提高,数字视频监控系统已逐渐取代传统的模拟式监控系统而广泛应用于银行、小区、宾馆等公共场所。在公共安全领域发挥着重要的作用。在未来的发展中,智能化监控系统将有广阔的应用前景。
本论文设计了一种基于运动目标识别技术的数字视频监控系统,对现场有无人员入侵进行判断,实现无人值守的智能化监控。论文探讨了一种基于SOPC(System on aProgrammable Chip,可编程片上系统)的视频监控系统的设计方案——CPU结合用户自定义逻辑的方案。此方案的设计思想是以下载到FPGA(Field Programmable Gate Array,片上可编程逻辑门阵列)的Nios5.0嵌入式软核CPU为系统控制模块,并运用FPGA逻辑单元实现视频图像处理。本方案中视频图像处理的大部分算法采用Verilog HDL(HardwareDescription Language,硬件描述语言)设计的图像处理模块实现,在很大程度上提高了系统速度。图象处理实现方面,考虑到在计算机视觉领域,人运动的视觉分析的研究具有广阔的应用前景,实时分割出运动的人体是研究起始的关键。在很多计算机视觉应用中,一个基础而关键的任务是从视频序列中确定运动目标,其中对于固定摄像机的监控视频运动目标的检测,最常用的方法是减背景技术。但由于构建背景模型需要考虑光照变化等很多因素,因此开发一个好的减背景算法面临很多挑战。为了使人们对该技术有个初步了解,本设计主要利用减背景技术实现运动目标检测的过程,最后指出了减背景技术的未来研究重点和发展方向。
智能化监控系统核心技术是运动目标检测,提取及通信技术。本文在熟悉相关资料的基础上基于NiosⅡ的智能化数字监控系统,研究了包括视频采集,图像数据的传输和运动目标检测,数据通信的软件编程和硬件系统集成等问题。在硬件系统集成过程中,系统采用专业CCD摄像头,并配以自定义组件实现其视频采集功能,在SOPC中配置集成系统。在软件编程中采用C语言,编写了运动目标提取的相应算法,并在此基础上构建了整个智能监控程序,包括视频图像采集,串口通信和运动目标提取等方面的编程。其中,提出了一种新的结合背景提取和差分的运动目标检测算法,可应用于静态背景下的运动目标分割。仿真结果表明该算法有效,可以提高运动目标检测的准确性,并有较好的实时性。
In recent years, with the rapid development of multimedia technical and the unceasing enhancement of computer performance, the digital video monitoring system has gradually substituted for traditional analog video monitoring system, which is widely applied to the public places such as back, and hotel and plays an important role in the public security domain. The intelligent video monitoring system will certainly have a broad application prospect in the future.
A digital video monitoring system based on moving object recognition is designed in this paper, which can judge whether there are some invaded people and realize intelligent monitoring system without attendant. A video identification system' s design plan based on SOPC (programmable system-on-chip), the plan of CPU integrating with user' s self-defining logic, is discussed in the thesis. Nios5. 0 embedded IP core CPU which has been downloaded to FPGA (Field Programmable Gate Array) acts as the control module of the system and FPGA' s logic units are used to implement the video image manipulation are the design thinking of this program. To a large extent, this plan improve the speed of the system, because most algorithm of the video image manipulation are implemented through using the video image manipulation modules of Verilog HDL(Hardware Description Language).Considered in the field of computer vision, the research about the analysis of the human motion have a wide application prospect and it is necessary to divide the moving human body timely in the realizing inspect of image manipulation. Identifying moving objects from a video sequence is a fundamental and critical task in many computer-vision applications. For surveillance video captured by static camera, the common approach is to perform background subtraction, which identifies moving objects from the portion of a video frame that differs significantly from a background model. However, there are many challenges in developing a good background subtraction algorithm for many factors such as changes in illumination should be considered in const. In order to enable the people to have a preliminary understanding about this technology, the design realized the process of identifying moving object by the technology of background subtraction, and pointed out the future research priorities and the direction of development about the technology of background subtraction at last.
The critical technology of intelligent video monitoring system is the moving-target detection, recognition and tracking. In the foundation of familiar with correlation data, the paper has realized one type of intelligent digital video monitoring system based on NIOSII, and has studied several problems including video frequency picture gathering, cradle head control, software programming of the moving-target detection and track and system integration. In the hardware system integration part, this system adopts specialized CCD camera and configuring user logic components so as to realizes the function of video gathering. And building system based on SOPC. In the software programming part, language C is used to construct the entire intelligent video frequency monitor system on the basis of moving-target recovery algorithm, including the programming of video frequency picture gathering, serial port communication and the detection track of mobile. These contains: Combined with background extraction and symmetry frame difference, a new moving-object detecting algorithm which can be used for moving-target segmentation under static background is proposed. The results of our experiment prove that the recommended algorithm is robust and effective. It can improve the accuracy of detecting moving objects and has good real-time performance. A novel algorithm of background initializing and updating applied to real-time video surveillance has been proposed in this paper.
本论文设计了一种基于运动目标识别技术的数字视频监控系统,对现场有无人员入侵进行判断,实现无人值守的智能化监控。论文探讨了一种基于SOPC(System on aProgrammable Chip,可编程片上系统)的视频监控系统的设计方案——CPU结合用户自定义逻辑的方案。此方案的设计思想是以下载到FPGA(Field Programmable Gate Array,片上可编程逻辑门阵列)的Nios5.0嵌入式软核CPU为系统控制模块,并运用FPGA逻辑单元实现视频图像处理。本方案中视频图像处理的大部分算法采用Verilog HDL(HardwareDescription Language,硬件描述语言)设计的图像处理模块实现,在很大程度上提高了系统速度。图象处理实现方面,考虑到在计算机视觉领域,人运动的视觉分析的研究具有广阔的应用前景,实时分割出运动的人体是研究起始的关键。在很多计算机视觉应用中,一个基础而关键的任务是从视频序列中确定运动目标,其中对于固定摄像机的监控视频运动目标的检测,最常用的方法是减背景技术。但由于构建背景模型需要考虑光照变化等很多因素,因此开发一个好的减背景算法面临很多挑战。为了使人们对该技术有个初步了解,本设计主要利用减背景技术实现运动目标检测的过程,最后指出了减背景技术的未来研究重点和发展方向。
智能化监控系统核心技术是运动目标检测,提取及通信技术。本文在熟悉相关资料的基础上基于NiosⅡ的智能化数字监控系统,研究了包括视频采集,图像数据的传输和运动目标检测,数据通信的软件编程和硬件系统集成等问题。在硬件系统集成过程中,系统采用专业CCD摄像头,并配以自定义组件实现其视频采集功能,在SOPC中配置集成系统。在软件编程中采用C语言,编写了运动目标提取的相应算法,并在此基础上构建了整个智能监控程序,包括视频图像采集,串口通信和运动目标提取等方面的编程。其中,提出了一种新的结合背景提取和差分的运动目标检测算法,可应用于静态背景下的运动目标分割。仿真结果表明该算法有效,可以提高运动目标检测的准确性,并有较好的实时性。
In recent years, with the rapid development of multimedia technical and the unceasing enhancement of computer performance, the digital video monitoring system has gradually substituted for traditional analog video monitoring system, which is widely applied to the public places such as back, and hotel and plays an important role in the public security domain. The intelligent video monitoring system will certainly have a broad application prospect in the future.
A digital video monitoring system based on moving object recognition is designed in this paper, which can judge whether there are some invaded people and realize intelligent monitoring system without attendant. A video identification system' s design plan based on SOPC (programmable system-on-chip), the plan of CPU integrating with user' s self-defining logic, is discussed in the thesis. Nios5. 0 embedded IP core CPU which has been downloaded to FPGA (Field Programmable Gate Array) acts as the control module of the system and FPGA' s logic units are used to implement the video image manipulation are the design thinking of this program. To a large extent, this plan improve the speed of the system, because most algorithm of the video image manipulation are implemented through using the video image manipulation modules of Verilog HDL(Hardware Description Language).Considered in the field of computer vision, the research about the analysis of the human motion have a wide application prospect and it is necessary to divide the moving human body timely in the realizing inspect of image manipulation. Identifying moving objects from a video sequence is a fundamental and critical task in many computer-vision applications. For surveillance video captured by static camera, the common approach is to perform background subtraction, which identifies moving objects from the portion of a video frame that differs significantly from a background model. However, there are many challenges in developing a good background subtraction algorithm for many factors such as changes in illumination should be considered in const. In order to enable the people to have a preliminary understanding about this technology, the design realized the process of identifying moving object by the technology of background subtraction, and pointed out the future research priorities and the direction of development about the technology of background subtraction at last.
The critical technology of intelligent video monitoring system is the moving-target detection, recognition and tracking. In the foundation of familiar with correlation data, the paper has realized one type of intelligent digital video monitoring system based on NIOSII, and has studied several problems including video frequency picture gathering, cradle head control, software programming of the moving-target detection and track and system integration. In the hardware system integration part, this system adopts specialized CCD camera and configuring user logic components so as to realizes the function of video gathering. And building system based on SOPC. In the software programming part, language C is used to construct the entire intelligent video frequency monitor system on the basis of moving-target recovery algorithm, including the programming of video frequency picture gathering, serial port communication and the detection track of mobile. These contains: Combined with background extraction and symmetry frame difference, a new moving-object detecting algorithm which can be used for moving-target segmentation under static background is proposed. The results of our experiment prove that the recommended algorithm is robust and effective. It can improve the accuracy of detecting moving objects and has good real-time performance. A novel algorithm of background initializing and updating applied to real-time video surveillance has been proposed in this paper.
引文
[1]何立民.嵌入式系统的定义与发展历史.北京:北京航空航天大学
[2]李华,杨军山.电视系统中的运动检测和运动估值.电视技术,1995(2)2-6
[3]王积分,张新荣.计算机图像识别.北京:中国铁道出版社1988
[4]张大波.嵌入式系统原理设计与应用.北京:机械工业出版社
[5]潘松,黄继业.SOPC技术实用教程。北京:清华大学出版社,2005年
[6]汪国强.SOPC技术与应用.北京:机械工业出版社,2006年
[7]彭澄廉.挑战SOC——基于Nios的SOPC设计与实践.北京:清华大学出版社,2004
[8]Michael D.Ciletti.Advanced Digital Design with the Verilog HDL.beijing:publishing House of Electronics Industry,2005
[9]Spartan-Ⅱ 2.5V FPGA Family:Complete DataSheet Product Specification.Philips Semiconductors.2004 Auf 2
[10]Real Time Image Processing based on FPGA Hardware Acceleration.Steffen klupsch,Markus Ernst research center Caesar center of advanced European studies and research Friedens platz 16,53111 Bonn
[11]Designing Digital Signal Processing with FPGAs.MentorGraphics.Feb 2003
[12]易克初,ffl斌,李刚强.FPGA设计中关键问题的研究.电子技术应用,2003年06期
[13]夏宇闻.Verilog数字系统设计教程.北京:北京航空航天大学出版社,2003年
[14]Jan M.Rabacy,Anantha Chandrakasan,Borivoie Nikolic.数字集成电路设计透视.北京:清华大学出版社,2004.
[15]薛宏熙,胡秀珠.计算机组成与设计.北京:清华大学出版社,2007年
[16]徐光辉,程东旭,黄如.基于FPGA的嵌入式开发与应用.北京:电子工业出版社,2006
[17]黄智伟.FPGA系统设计与实践.北京:电子工业出版社,2005年
[18]易国华.灰度图象序列中运动目标提取方法的研究.鄂州大学学报,2003,Vol.10 No.4,25-27
[19]刘永,戴礼容,宋彦.一种静态背景下的运动目标提取算法.中国图像学报,2003,7:760-163
[20]周新伦,柳健,刘华志.数字图像处理.北京:国防工业出版社,1986
[21]李金宗,李宁宁.图像积累与自适应门限检测技术的研究.电子学报 1997,25
[22]阮秋琦.数字图像处理学.北京:电子工业出版社,2001年.
[23]李彬,刘冀伟,韩鸿哲,李正熙.复杂背景下人体骨架的提取[J].微计算机信息,2004,20(7):43-44.
[24]李春明,李玉山,张大朴,刘洋.运动人物的检测、跟踪与识别综述[DB/OL].http://www.paper.edu.cn,2005.04-18.
[25]Neff A,Colonnese S,Russo G,Talone P.Automatic moving object and background separation[J].Signal Processing,1998,66(2):219-232.
[26]Me chR,WolbomM.A noise robust method for 2D shape esti.marion of moving object in video sequences comidering alnoving camera[J].Signal Processing,1998,66(2):203-205.
[27]A Shio,J Sklansky.Segmentation of people in motion.Proc.IEEE Workshop on Visual Motion[C].1991.325-332.
[28]张建荣,姜昱明.实时跟踪系统中运动人体图像分割[J].计算机仿真,2004,21(6):54-56.
[29]唐常青,吕宏伯.数学形态学方法及其应用.北京:科学出版社,1999年.
[30]龚炜,石青云,程民德.数字空间中的数学形态学——理论及应用.北京:科学出版社,1997年.
[31]乔晓丹,张鹏.一个基于Linux操作系统的嵌入式网关的实现[J]微计算机信息,22005.21,7-2.
[32]朱刚等,Linux网络编程,科学出版社,2000.
[33]戴青云,余英林.数学形态学在图像处理中的应用进展[J].控制理论应用,2001,8(4):478-482.
[34]Hochong Park,Roland T Chin.Decopmosition of arbitrarily shaped morphological structuring elements[J].IEEE Trans on PAMI,1991,7(1):2-15.
[35]龚炜.数学形态学中结构元素的分解[J]高校应用数学学报,1998(3):340-349.
[36]R.Brunelli,O.Mich,C.M.Modena,A survey on the automatic indexing of video data,J.Visual Commun.Image Representation 10(1999)78-112.
[37]R.Koenen,F.Pereira,L.Chiariglione,MPEG-4:context and objectives,Signal Process.Image Commun.9(4)(1997)295-304.
[38]J.Martinez,R.Koenen,F.Pereira,MPEG-7:the generic multimediacontent description standard, part 1, multimedia, IEEE 9 (2002)78-87.
[39] D.S. Zhang, G Lu, Segmentation of moving objects in image sequence: a review, Circuits Syst. Signal Process. 20(2) (2001)143-183.
[40] P.L. Rosin, Thresholding for change detection, in: Proceedings of the Sixth International Conference on Computer Vision. IEEE Computer Society, Silver Spring,MD,1998,274p.
[41] P. Giaccone, G. Jones, Segmentation of global motion using temporal probabilistic classification, in: British Machine Vision Conference, 1998, pp. 619-628.
[42] H. Sawhney, S. Ayer, Compact representations of videos through dominant and multiple motion estimation, IEEE Trans. Pattern Analy.Mach. Intell.18 (8) (1996) 814-830.
[43] M. Irani, P. Anandan, J. Bergen, R. Kumar, S. Hsu, Efficient representations of video sequences and their applications, Signal Process. Image Commun.8(4) (1996) 327-351.
[44] S. Park, J. Aggarwal, Segmentation and tracking of interacting humanbody parts under occlusion and shadowing, in: IEEE Workshop on Motion and Video Computing, 2002, pp. 105-111.
[45] I. Cohen, G Medioni, Detecting and tracking moving objects in video surveil lance, in: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 1999, pp. II: 319-325.
[2]李华,杨军山.电视系统中的运动检测和运动估值.电视技术,1995(2)2-6
[3]王积分,张新荣.计算机图像识别.北京:中国铁道出版社1988
[4]张大波.嵌入式系统原理设计与应用.北京:机械工业出版社
[5]潘松,黄继业.SOPC技术实用教程。北京:清华大学出版社,2005年
[6]汪国强.SOPC技术与应用.北京:机械工业出版社,2006年
[7]彭澄廉.挑战SOC——基于Nios的SOPC设计与实践.北京:清华大学出版社,2004
[8]Michael D.Ciletti.Advanced Digital Design with the Verilog HDL.beijing:publishing House of Electronics Industry,2005
[9]Spartan-Ⅱ 2.5V FPGA Family:Complete DataSheet Product Specification.Philips Semiconductors.2004 Auf 2
[10]Real Time Image Processing based on FPGA Hardware Acceleration.Steffen klupsch,Markus Ernst research center Caesar center of advanced European studies and research Friedens platz 16,53111 Bonn
[11]Designing Digital Signal Processing with FPGAs.MentorGraphics.Feb 2003
[12]易克初,ffl斌,李刚强.FPGA设计中关键问题的研究.电子技术应用,2003年06期
[13]夏宇闻.Verilog数字系统设计教程.北京:北京航空航天大学出版社,2003年
[14]Jan M.Rabacy,Anantha Chandrakasan,Borivoie Nikolic.数字集成电路设计透视.北京:清华大学出版社,2004.
[15]薛宏熙,胡秀珠.计算机组成与设计.北京:清华大学出版社,2007年
[16]徐光辉,程东旭,黄如.基于FPGA的嵌入式开发与应用.北京:电子工业出版社,2006
[17]黄智伟.FPGA系统设计与实践.北京:电子工业出版社,2005年
[18]易国华.灰度图象序列中运动目标提取方法的研究.鄂州大学学报,2003,Vol.10 No.4,25-27
[19]刘永,戴礼容,宋彦.一种静态背景下的运动目标提取算法.中国图像学报,2003,7:760-163
[20]周新伦,柳健,刘华志.数字图像处理.北京:国防工业出版社,1986
[21]李金宗,李宁宁.图像积累与自适应门限检测技术的研究.电子学报 1997,25
[22]阮秋琦.数字图像处理学.北京:电子工业出版社,2001年.
[23]李彬,刘冀伟,韩鸿哲,李正熙.复杂背景下人体骨架的提取[J].微计算机信息,2004,20(7):43-44.
[24]李春明,李玉山,张大朴,刘洋.运动人物的检测、跟踪与识别综述[DB/OL].http://www.paper.edu.cn,2005.04-18.
[25]Neff A,Colonnese S,Russo G,Talone P.Automatic moving object and background separation[J].Signal Processing,1998,66(2):219-232.
[26]Me chR,WolbomM.A noise robust method for 2D shape esti.marion of moving object in video sequences comidering alnoving camera[J].Signal Processing,1998,66(2):203-205.
[27]A Shio,J Sklansky.Segmentation of people in motion.Proc.IEEE Workshop on Visual Motion[C].1991.325-332.
[28]张建荣,姜昱明.实时跟踪系统中运动人体图像分割[J].计算机仿真,2004,21(6):54-56.
[29]唐常青,吕宏伯.数学形态学方法及其应用.北京:科学出版社,1999年.
[30]龚炜,石青云,程民德.数字空间中的数学形态学——理论及应用.北京:科学出版社,1997年.
[31]乔晓丹,张鹏.一个基于Linux操作系统的嵌入式网关的实现[J]微计算机信息,22005.21,7-2.
[32]朱刚等,Linux网络编程,科学出版社,2000.
[33]戴青云,余英林.数学形态学在图像处理中的应用进展[J].控制理论应用,2001,8(4):478-482.
[34]Hochong Park,Roland T Chin.Decopmosition of arbitrarily shaped morphological structuring elements[J].IEEE Trans on PAMI,1991,7(1):2-15.
[35]龚炜.数学形态学中结构元素的分解[J]高校应用数学学报,1998(3):340-349.
[36]R.Brunelli,O.Mich,C.M.Modena,A survey on the automatic indexing of video data,J.Visual Commun.Image Representation 10(1999)78-112.
[37]R.Koenen,F.Pereira,L.Chiariglione,MPEG-4:context and objectives,Signal Process.Image Commun.9(4)(1997)295-304.
[38]J.Martinez,R.Koenen,F.Pereira,MPEG-7:the generic multimediacontent description standard, part 1, multimedia, IEEE 9 (2002)78-87.
[39] D.S. Zhang, G Lu, Segmentation of moving objects in image sequence: a review, Circuits Syst. Signal Process. 20(2) (2001)143-183.
[40] P.L. Rosin, Thresholding for change detection, in: Proceedings of the Sixth International Conference on Computer Vision. IEEE Computer Society, Silver Spring,MD,1998,274p.
[41] P. Giaccone, G. Jones, Segmentation of global motion using temporal probabilistic classification, in: British Machine Vision Conference, 1998, pp. 619-628.
[42] H. Sawhney, S. Ayer, Compact representations of videos through dominant and multiple motion estimation, IEEE Trans. Pattern Analy.Mach. Intell.18 (8) (1996) 814-830.
[43] M. Irani, P. Anandan, J. Bergen, R. Kumar, S. Hsu, Efficient representations of video sequences and their applications, Signal Process. Image Commun.8(4) (1996) 327-351.
[44] S. Park, J. Aggarwal, Segmentation and tracking of interacting humanbody parts under occlusion and shadowing, in: IEEE Workshop on Motion and Video Computing, 2002, pp. 105-111.
[45] I. Cohen, G Medioni, Detecting and tracking moving objects in video surveil lance, in: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 1999, pp. II: 319-325.