视频监控系统的设计和实现
|
摘要
随着计算机信息处理技术、视觉处理技术、多媒体技术的快速发展,计算机视频监控系统在国家安全方面所起到的重要作用越来越明显,受到了各界人士的广泛关注。而视频监控技术中运动目标检测技术是这个领域的重要研究方向,本论文通过对现有主流目标检测算法的研究,在总结和分析的基础上提出一种新的检测算法,能够实现背景有效更新,并应用于一个完整的智能视频监控系统中,使系统具备更好的检测效果。
运动目标检测技术是计算机视频监控系统中首要的和关键的技术,它是以后计算机视觉技术和监控技术发展到更高级水平的前提。运动目标检测的目的是为了找到产生运动的区域,就是要实现在视频图像序列中提取真实监控场景中的前景运动目标,并对检测出来的运动目标进行分割。
本论文首先对目前主流的视频检测算法,如光流法、背景差分法、帧间差分法、熵检测法进行实验分析与研究,比较优势与不足、适用范围、存在的问题及处理办法。在掌握背景差分法和帧间差分法的原理的基础上,进行结合,提出了使用于摄像机固定的运动目标检测方法,能够有效的进行背景更新,更精确的提取出运动目标,并能减轻由于外界光照条件、背景扰动、噪音等干扰因素造成的影响。试验表明,该算法准确,有很好的检测效果。
With the computer information processing, visual processing, the rapid development of multimedia technology, computer video surveillance system in terms of national security, more and more obvious the important role played by the public attention. The video monitoring technology in motion detection technology is an important research direction in this field, this paper through the existing main target detection algorithm, the summary and analysis based on the a new detection algorithm, can achieve effective update the background and applied to a complete intelligent video surveillance system, the system has better detection.
Moving target detection technology is the computer video surveillance system in the first and key technologies, it is the future of computer vision technology and monitoring technology to a more advanced level of development of the premise. The purpose of moving target detection is to find areas in motion, is to achieve the real video image sequences to extract the prospect of monitoring moving objects in the scene, and the detected moving object segmentation.
In this thesis, the current mainstream video detection algorithms, such as optical flow, background subtraction method, frame difference method, entropy detection analysis and research experiments to compare the advantages and disadvantages, scope, problems and approaches. In the control background and frame difference method difference method based on the principle of conduct, proposed the use of video cameras fixed in the moving target detection methods, background can be effectively updated, more accurate extraction of the moving target, and to reduce as external light conditions, background disturbance, noise and other interference factors impact. Tests show that the algorithm accurately, a very good test results.
运动目标检测技术是计算机视频监控系统中首要的和关键的技术,它是以后计算机视觉技术和监控技术发展到更高级水平的前提。运动目标检测的目的是为了找到产生运动的区域,就是要实现在视频图像序列中提取真实监控场景中的前景运动目标,并对检测出来的运动目标进行分割。
本论文首先对目前主流的视频检测算法,如光流法、背景差分法、帧间差分法、熵检测法进行实验分析与研究,比较优势与不足、适用范围、存在的问题及处理办法。在掌握背景差分法和帧间差分法的原理的基础上,进行结合,提出了使用于摄像机固定的运动目标检测方法,能够有效的进行背景更新,更精确的提取出运动目标,并能减轻由于外界光照条件、背景扰动、噪音等干扰因素造成的影响。试验表明,该算法准确,有很好的检测效果。
With the computer information processing, visual processing, the rapid development of multimedia technology, computer video surveillance system in terms of national security, more and more obvious the important role played by the public attention. The video monitoring technology in motion detection technology is an important research direction in this field, this paper through the existing main target detection algorithm, the summary and analysis based on the a new detection algorithm, can achieve effective update the background and applied to a complete intelligent video surveillance system, the system has better detection.
Moving target detection technology is the computer video surveillance system in the first and key technologies, it is the future of computer vision technology and monitoring technology to a more advanced level of development of the premise. The purpose of moving target detection is to find areas in motion, is to achieve the real video image sequences to extract the prospect of monitoring moving objects in the scene, and the detected moving object segmentation.
In this thesis, the current mainstream video detection algorithms, such as optical flow, background subtraction method, frame difference method, entropy detection analysis and research experiments to compare the advantages and disadvantages, scope, problems and approaches. In the control background and frame difference method difference method based on the principle of conduct, proposed the use of video cameras fixed in the moving target detection methods, background can be effectively updated, more accurate extraction of the moving target, and to reduce as external light conditions, background disturbance, noise and other interference factors impact. Tests show that the algorithm accurately, a very good test results.
引文
[1]代科学,李国辉,涂丹等.。监控视频运动目标检测背景技术的研究现状与展望[J]。中国图像图形学报,2006,07:919-927
[2]宋磊等。视频监控系统概述[J]。测控技术,2003,5:33-35
[3]王莹莹。智能视频监控粉墨登场。中国计算机用户,2006,10:40-41
[4] R Collins, A Lipton et al. A system for video surveillance and monitoring: VSAM final report. CMU-RI-TR-00-I2, Robotic Institute Carnegie Mellon University, 2000.
[5] Remagnino P, Tan T, Baker K. Mufti-agent visual surveillance of dynamic scenes. Image and Vision Computing, 1998,16(8):529-532.
[6] R Collins, A. Lipton, and T. Kaasade. Introduction to the special section on video surveillance IEEE Trans. Patern Analysis and Machine Intelligence. 22(8):745-746, August 2000.
[7] Jwu-Sheng Hu, Tzung-Min Su and Shr-Chi Jeng,“Robust Backgroud Subtraction with Shadow and Highlight Removal for Indoor Surveillance,”Proceedings of the 2006 IEEE/RSJ International Conference on Intelligent Robots and Systems October 9-15,2006,Beijing, China,4545-4551.
[8]朱铮涛等。基于图像嫡的自动聚焦函数研究。光学与精密工程。2004,5:537-542
[9]危水根等。一种基于时间差分运动检测的改进方法。南昌航空工业学院学报。2005,3:15-19
[10]代科学等。监控视频运动目标检测减背景技术的研究现状和展望。中国图像图形学报。2 006,7:919-927
[11]谢攀等。一种基于尺度变化的DCT自动聚焦算法。华大学学报。2003,1:55-58
[12]李振华。图像通信与监控系统。北京:中国铁道出版社,1994
[13]曹莉华,胡晓峰。视频压缩编码标准的进展。世界网络与多媒体,2000,7:18-21
[14] A. Aydin Alatan, Levent Onural., Michael Wollborn, et al. Image Sequence Analysis for Emerging Interactive Multimedia Services. The European COST 211 Framework[J]. IEEE Transactions on Circuits and Systems for Video Technology. 1998-11, (8,7):802-813.
[15]魏志强,纪筱鹏,冯业伟。基于自适应背景图像更新的运动目标检测方法[J]。电子学报,2005,33 (12):2261-2264。
[16] WEI Zhiqiang , Ji Xiaopeng , FENG Yewei . A moving object detection method based on selfadaptive updating of background[J]. Acta Electronica Sinica, 2005, 33 (12): 2261-2264.
[17]朱辉,李在铭。视频序列中运动目标检测技术。信号处理,2002,5:448-451
[18]李俊韬,张海,范跃祖,王力。复杂场景条件下的运动目标检测方法。光电工程,2004,31:36-39
[19] Elgammal A, Hawood D, Davis L S. Non-Parametric model for background subtraction Proc. ECCV, 2000, 7:751-767.
[20] Richard P. Wildes. A measure of motion salience for surveillance application. In Proc of Image Processing, 1998.
[21]郝志成,朱明。智能目标检测与跟踪系统的设计与实现[J]。光电工程,2007,34(1):27-3
[22]王陈阳,周明全,耿国华。基于自适应背景模型运动目标检测[J]。计算机技术与发展,2007,17(4):21-23
[23]李晴,徐群。复杂场景下多运动目标速度检测技术的实现[J]。计算机与数字工程,2006,34(11):167-171
[24] Chang C J, HSIEH J W, CHEN Y S, et al. Tracking multiple moving objects using a level-set method[J]. International Journal of Pattern Recognition and Artificial Intelligence, 2004, 18(2): 101-125.
[25] Parameswaran V, Chellappa, R. Quasi-invariants for human action representation and recognition[A]. In Proc ICPR[C]. Quebec: IEEE Computer Society Press, 2002, 307-310.
[26] Bobick A F, Davis J W. The recognition of human movement using temporal templates[J]. IEEE Trans. PAMI, 2001,15(3): 257-267.
[27]王哲,常发亮。一种基于立体视觉的运动目标检测算法[J]。计算机应用,2006,26(11)2724-2726
[28]郁梅,蒋刚毅。智能交通系统中的计算机视觉技术应用[J]。计算机工程及应用,2001,37(10):101-103
[29] Wang F. and Zhang, Y.. Improving TCP Performance over Mobile Ad-Hoc Networks with Out-of-Order Detection and Response, Proceeding of ACM MOBIHOC, 2002.
[30]秦涛,周泽魁。一种检测序列图像中运动目标的新方法[J]。计算机应用与软件。2004,21(9):105-107
[31]何卫华,李平,文玉海等。复杂背景下基于图像融合的运动目标轮廓提起[J]。计算机应用。2006,26(1):123-126
[32]单勇。复杂条件下视频运动检测和跟踪[D]。博士学位论文,国防科技大学,长沙,2006
[33]张泽旭,李金宗,李宁宁,基于光流场分割和Canny迈缘提取融合算法的运动目标检测。电子学报,2003,31(9):1299-1302
[34]魏波。点时空约束图像目标跟踪理论与实时实现技术研究[D]。成都:电子科技大学,2000
[35]付萍,方帅,徐心和等,视频监控系统中运动目标检测的阴影去除方法。计算机工程,2007,33(10):22-24
[2]宋磊等。视频监控系统概述[J]。测控技术,2003,5:33-35
[3]王莹莹。智能视频监控粉墨登场。中国计算机用户,2006,10:40-41
[4] R Collins, A Lipton et al. A system for video surveillance and monitoring: VSAM final report. CMU-RI-TR-00-I2, Robotic Institute Carnegie Mellon University, 2000.
[5] Remagnino P, Tan T, Baker K. Mufti-agent visual surveillance of dynamic scenes. Image and Vision Computing, 1998,16(8):529-532.
[6] R Collins, A. Lipton, and T. Kaasade. Introduction to the special section on video surveillance IEEE Trans. Patern Analysis and Machine Intelligence. 22(8):745-746, August 2000.
[7] Jwu-Sheng Hu, Tzung-Min Su and Shr-Chi Jeng,“Robust Backgroud Subtraction with Shadow and Highlight Removal for Indoor Surveillance,”Proceedings of the 2006 IEEE/RSJ International Conference on Intelligent Robots and Systems October 9-15,2006,Beijing, China,4545-4551.
[8]朱铮涛等。基于图像嫡的自动聚焦函数研究。光学与精密工程。2004,5:537-542
[9]危水根等。一种基于时间差分运动检测的改进方法。南昌航空工业学院学报。2005,3:15-19
[10]代科学等。监控视频运动目标检测减背景技术的研究现状和展望。中国图像图形学报。2 006,7:919-927
[11]谢攀等。一种基于尺度变化的DCT自动聚焦算法。华大学学报。2003,1:55-58
[12]李振华。图像通信与监控系统。北京:中国铁道出版社,1994
[13]曹莉华,胡晓峰。视频压缩编码标准的进展。世界网络与多媒体,2000,7:18-21
[14] A. Aydin Alatan, Levent Onural., Michael Wollborn, et al. Image Sequence Analysis for Emerging Interactive Multimedia Services. The European COST 211 Framework[J]. IEEE Transactions on Circuits and Systems for Video Technology. 1998-11, (8,7):802-813.
[15]魏志强,纪筱鹏,冯业伟。基于自适应背景图像更新的运动目标检测方法[J]。电子学报,2005,33 (12):2261-2264。
[16] WEI Zhiqiang , Ji Xiaopeng , FENG Yewei . A moving object detection method based on selfadaptive updating of background[J]. Acta Electronica Sinica, 2005, 33 (12): 2261-2264.
[17]朱辉,李在铭。视频序列中运动目标检测技术。信号处理,2002,5:448-451
[18]李俊韬,张海,范跃祖,王力。复杂场景条件下的运动目标检测方法。光电工程,2004,31:36-39
[19] Elgammal A, Hawood D, Davis L S. Non-Parametric model for background subtraction Proc. ECCV, 2000, 7:751-767.
[20] Richard P. Wildes. A measure of motion salience for surveillance application. In Proc of Image Processing, 1998.
[21]郝志成,朱明。智能目标检测与跟踪系统的设计与实现[J]。光电工程,2007,34(1):27-3
[22]王陈阳,周明全,耿国华。基于自适应背景模型运动目标检测[J]。计算机技术与发展,2007,17(4):21-23
[23]李晴,徐群。复杂场景下多运动目标速度检测技术的实现[J]。计算机与数字工程,2006,34(11):167-171
[24] Chang C J, HSIEH J W, CHEN Y S, et al. Tracking multiple moving objects using a level-set method[J]. International Journal of Pattern Recognition and Artificial Intelligence, 2004, 18(2): 101-125.
[25] Parameswaran V, Chellappa, R. Quasi-invariants for human action representation and recognition[A]. In Proc ICPR[C]. Quebec: IEEE Computer Society Press, 2002, 307-310.
[26] Bobick A F, Davis J W. The recognition of human movement using temporal templates[J]. IEEE Trans. PAMI, 2001,15(3): 257-267.
[27]王哲,常发亮。一种基于立体视觉的运动目标检测算法[J]。计算机应用,2006,26(11)2724-2726
[28]郁梅,蒋刚毅。智能交通系统中的计算机视觉技术应用[J]。计算机工程及应用,2001,37(10):101-103
[29] Wang F. and Zhang, Y.. Improving TCP Performance over Mobile Ad-Hoc Networks with Out-of-Order Detection and Response, Proceeding of ACM MOBIHOC, 2002.
[30]秦涛,周泽魁。一种检测序列图像中运动目标的新方法[J]。计算机应用与软件。2004,21(9):105-107
[31]何卫华,李平,文玉海等。复杂背景下基于图像融合的运动目标轮廓提起[J]。计算机应用。2006,26(1):123-126
[32]单勇。复杂条件下视频运动检测和跟踪[D]。博士学位论文,国防科技大学,长沙,2006
[33]张泽旭,李金宗,李宁宁,基于光流场分割和Canny迈缘提取融合算法的运动目标检测。电子学报,2003,31(9):1299-1302
[34]魏波。点时空约束图像目标跟踪理论与实时实现技术研究[D]。成都:电子科技大学,2000
[35]付萍,方帅,徐心和等,视频监控系统中运动目标检测的阴影去除方法。计算机工程,2007,33(10):22-24