视频监控中的运动对象分割技术研究
|
摘要
视频监控技术是计算机视觉领域一个新兴的应用方向和备受关注的前沿课题,是计算机科学、机器视觉、图像工程、模式识别和人工智能等多种学科的结晶,广泛应用于城市道路交通监控,安防监控等各方面。
运动对象分割是视频监控中的一项关键技术,分割的准确性直接影响后续任务的有效性,因此具有十分重要的意义。目前,对视频对象的分割有很多优秀的算法,但这些算法都是针对特定的应用提出来的,而适合任何场景的全自动视频分割算法仍然是一个有待解决的经典难题。本文基于LK光流法和混合高斯模型的视频对象分割算法进行了深入研究和大量实验,得到了一系列有价值的结论和研究成果。本文主要工作可总结为以下几个方面:
1.分析现有光流法的性能及优缺点,提出了一种改进的算法。本文运用Gaussian金字塔降低运动对象的速度,使用结合Gaussian分布的LK光流法进行分割,得到运动对象。该算法提高了实时性,能获得更加精确的运动对象。
2.考虑到初始背景帧的提取好坏直接影响到背景建模的性能,在现有的初始背景帧的提取方法基础上,提出了一种MEAMO方法来产生初始的背景帧,减少了初始误差,有利于后续任务的完成。
3.由于阴影对运动对象的分割会产生严重的影响,提出了一种基于混合高斯模型的自适应阴影检测算法。该算法选择在CIE LUV颜色空间,利用亮度分量L对背景建模,再利用高斯分布对前景和背景中的L分量比值进行自适应阴影检测。该算法实现了自适应的阴影检测,具有较强的鲁棒性和较高的分割精度。
总之,本文对视频监控中的分割算法做了进一步的研究,通过实验证明,取得了满意的实验结果。
Video surveillance technology is the new-emerging application direction in the filed of Computer Vision and is attracting more and more attention.It spans many subjects including computer science,machine vision,image engineering,pattern recognition, artificial intelligence and so on.It widely applies in city road traffic surveillance and security surveillance.
The movement objects segmentation is the key technology in video surveillance.The veracity of segmentation directly affects the effectiveness of latter tasks.So it has very important meaning.At present,there have many excellent segmentation algorithms for video objects,but these algorithms are proposed in special application environment.The automatic video segmentation algorithm that can use in any scene is still a classical difficulty need to be solved.The video objects segmentation algorithms are deeply studied based on LK Optical Flow and Mixture Gaussian Model in this thesis.Through a great deal of experiments,the thesis acquired a series of valuable results which can be summarized in the following aspects:
1.On the basis of analyzing the performance of existing Optical Flow methods,an improved algorithm is presented.This thesis uses Gaussian pyramid reduce the speed of movement objects,and then segments pictures use Gaussian distribution combine with LK Optical Flow.The algorithm improves the real-time capability and can get more accurate movement objects.
2.Considering the extraction effects of initial background frame directly impact the performance of background modeling,an MEAMO algorithm of producing initial background frame is proposed based on existing methods.The algorithm can decrease the initial error and it is helpful to achieve the latter tasks.
3.For the shadows produce serious impacts to movement objects segmentation,an adaptive shadows detection algorithm based on Mixture Gaussian Model is presented. Choosing CIE LUV color space,the paper uses L weight do background modeling,and then uses Gaussian distribution do adaptive shadows detection for the ratio of foreground's L weight and background's L weight.The algorithm achieves adaptive shadows detection.It has strong robustness and high accuracy.
In conclusion,this thesis makes farther research for the segmentation algorithms in video surveillance.The experimental results of each researched algorithm are intending and good.
运动对象分割是视频监控中的一项关键技术,分割的准确性直接影响后续任务的有效性,因此具有十分重要的意义。目前,对视频对象的分割有很多优秀的算法,但这些算法都是针对特定的应用提出来的,而适合任何场景的全自动视频分割算法仍然是一个有待解决的经典难题。本文基于LK光流法和混合高斯模型的视频对象分割算法进行了深入研究和大量实验,得到了一系列有价值的结论和研究成果。本文主要工作可总结为以下几个方面:
1.分析现有光流法的性能及优缺点,提出了一种改进的算法。本文运用Gaussian金字塔降低运动对象的速度,使用结合Gaussian分布的LK光流法进行分割,得到运动对象。该算法提高了实时性,能获得更加精确的运动对象。
2.考虑到初始背景帧的提取好坏直接影响到背景建模的性能,在现有的初始背景帧的提取方法基础上,提出了一种MEAMO方法来产生初始的背景帧,减少了初始误差,有利于后续任务的完成。
3.由于阴影对运动对象的分割会产生严重的影响,提出了一种基于混合高斯模型的自适应阴影检测算法。该算法选择在CIE LUV颜色空间,利用亮度分量L对背景建模,再利用高斯分布对前景和背景中的L分量比值进行自适应阴影检测。该算法实现了自适应的阴影检测,具有较强的鲁棒性和较高的分割精度。
总之,本文对视频监控中的分割算法做了进一步的研究,通过实验证明,取得了满意的实验结果。
Video surveillance technology is the new-emerging application direction in the filed of Computer Vision and is attracting more and more attention.It spans many subjects including computer science,machine vision,image engineering,pattern recognition, artificial intelligence and so on.It widely applies in city road traffic surveillance and security surveillance.
The movement objects segmentation is the key technology in video surveillance.The veracity of segmentation directly affects the effectiveness of latter tasks.So it has very important meaning.At present,there have many excellent segmentation algorithms for video objects,but these algorithms are proposed in special application environment.The automatic video segmentation algorithm that can use in any scene is still a classical difficulty need to be solved.The video objects segmentation algorithms are deeply studied based on LK Optical Flow and Mixture Gaussian Model in this thesis.Through a great deal of experiments,the thesis acquired a series of valuable results which can be summarized in the following aspects:
1.On the basis of analyzing the performance of existing Optical Flow methods,an improved algorithm is presented.This thesis uses Gaussian pyramid reduce the speed of movement objects,and then segments pictures use Gaussian distribution combine with LK Optical Flow.The algorithm improves the real-time capability and can get more accurate movement objects.
2.Considering the extraction effects of initial background frame directly impact the performance of background modeling,an MEAMO algorithm of producing initial background frame is proposed based on existing methods.The algorithm can decrease the initial error and it is helpful to achieve the latter tasks.
3.For the shadows produce serious impacts to movement objects segmentation,an adaptive shadows detection algorithm based on Mixture Gaussian Model is presented. Choosing CIE LUV color space,the paper uses L weight do background modeling,and then uses Gaussian distribution do adaptive shadows detection for the ratio of foreground's L weight and background's L weight.The algorithm achieves adaptive shadows detection.It has strong robustness and high accuracy.
In conclusion,this thesis makes farther research for the segmentation algorithms in video surveillance.The experimental results of each researched algorithm are intending and good.
引文
[1]严勇.高速公路隧道视频监控系统的算法研究及实现[D].中国优秀硕士学位论文全文数据库,重庆大学,2006.
[2]刘怀强.基于视频的车辆检测与跟踪技术研究[D].中国优秀硕士学位论文全文数库,中国海洋大学,2007.
[3]王素玉,沈兰荪.智能视觉监控技术研究进展[J].中国图象图形学报,2007,12(9):1505-1513.
[4]Haritaoglu I,Harwood D,Davis L.W4:Real-time surveillance of people and their activities[J].IEEE Trans Pattern Analysis and Machine Intelligence,2000,22(8):809-830.
[5]S.J.McKenna,S.Jabri,Z.Duric,A.Rosenfeld.Tracking groups of people[J].Computer Vision and Image Understanding,2000,80(1):42-56.
[6]Stauffer C,Grimson W.Adaptive background mixture models for real-time tracking[C].In:Proceeding of IEEE Conference on Computer Vision and Pattern Recognition,Fort Colline,Colorado,USA,1999:246-252.
[7]徐东彬,刘昌平,黄磊.基于概率统计自适应背景模型的运动目标检测方法[J].中国图象图形学报,2008,13(2):351-358.
[8]万峻甫,刘建伟,向怀坤,曹泉,王钧生.交通视频序列阴影检测算法研究[J].中国图象图形学报,2008,13(3):468-471.
[9]王成儒,顾广华.基于差分交集的视频对象分割与跟跟踪算法[J].光学技术,2004,30(5):564-566.
[10]倪峰.基于变化模板检测的视频对象分割算法研究[D].中国优秀硕士学位论文全文数据库,苏州大学,2007.
[11]石敏,易清明,刘金梅.一种基于边缘检测的去块效应算法[J].计算机工程与应用,2007,43(3):27-29.
[12]龚声蓉,刘纯平,王强.数字图像处理与分析[M].北京:清华大学出版社,2007.
[13]田宏阳,陈辉,马文静.一种动态场景下基于时空信息的视频对象提取算法[J].中 国图象图形学报,2007,12(9):1652-1657.
[14]杨文明.时空联合的视频对象分割[D].中国优秀硕士学位论文全文数数库,浙江大学,2006.
[15]高文,陈熙霖.计算机视觉--算法与系统原理[M].北京:清华大学出版社,1999.
[16]E P Ong.Robust optical flow computation based on least median of squares regression [J].Computer Vision,1999,31(1):51-82.
[17]邓玉春,姜昱明,张建荣.视频序列图像中运动对象分割综述[J].计算机应用研究,2005,1:8-11.
[18]Wollbom M,Mech R.Refined procedure for objective evaluation of video segmentation algorithm.Doc.ISO/IEC JTC 1/SC29/WG11 M3448,March,1998.
[19]J.L.Barron,D.J.Fleet,S.S.Beauchemin.Performance of optical techniques[J].International Journal of Computer Vision,1994,12(1):43-77.
[20]Simon Denman,Vinod Chandran,Sridha Sridharan.An adaptive optical flow technique for person tracking systems[J].Pattern Recognition Letters,2007,28(10):1232-1239.
[21]E.Jayabalan,Dr.A.Krishnan,R.Pugazendi.Non rigid object tracking in aerial videos by combined snaked and optical flow technique[J].Computer Graphics,Imaging and Visualization,2007,21(6):388-396.
[22]吴新根,罗立民.一种改进的光流场计算方法[J].电子学报,2000,28(1):130-132.
[23]Nils Papenberg.Highly accurate optic flow computation with theoretically justified warping[J].International Journal of Computer Vision,2006,67(2):141-158.
[24]杨国亮,王志良,牟世堂,解仑,刘冀伟.一种改进的光流算法[J].计算机工程,2006,32(15):187-189.
[25]曾浩,高秀娟,曾孝平.快速运动估计中一种改进的块匹配免疫算法[J].计算机应用,2008,28(8):2147-2149.
[26]洪波,余松煜.基于对象的菱形搜索运动估计方法[J].数据采集与处理,2001,16(1):110-114.
[27]Gong Sheng-rong,Zhou xiang.HDS:A fast and hierarchical diamond search algorithm in video motion estimation[C].Proceedings of SPIE - The International Society for Optical Engineering,v 6044,MIPPR 2005:Image Analysis Techniques,2005:60440D-1-60440D-8.
[28]李炜,周兵,李波.运动适量场自适应搜索算法[J].计算机学报,2003,23(2):168-173.
[29]章毓晋.图像工程(上册)图像处理[M].北京:清华大学出版社,2007.
[30]肖梅,韩崇昭,张雷.一种视频序列的背景提取算法[J].光电工程,2005,32(4):78-80.
[31]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,2007,12(9):1585-1589.
[32]张海青,李厚强.基于均匀混合模型的运动检测[J].计算机工程与应用,2007,43(9):28-30.
[33]Peng Suo,Yanjiang Wang.An improved adaptive background modeling algorithm based on Gaussian Mixture Model[C].In:Proceeding of the 9th International Conference on Signal Processing,2008:1436-1439.
[34]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.
[35]曾艳,于濂.一种新的道路交通背景提取算法及研究[J].中国图象图形学报,2008,13(3):593-599.
[36]杜友田,陈峰,徐文立.基于区域的运动阴影检测方法[J].清华大学学报,2006,46(1):141-144.
[37]万峻甫,刘建伟,向怀坤,曹泉,王钧生.交通视频序列阴影检测算法研究[J].中国图象图形学报,2008,13(3):467-471.
[38]Cavallaro A,Salvador E,Ebrahimi T.Shadow-aware object-based video processing[C].IEE Proceedings-Vision Image Signal Process,2005,152(4):398-406.
[39]陈柏生,陈锻生.基于归一化RGB彩色模型的运动阴影检测[J].计算机应用,2006,26(8):1879-1881.
[40]Siala.k,Chakchouk.M,Chaieb.F,Besbes.O.Moving shadows detection with support vector domain description in the color ratios space[C].In:Proceedings of the 17th International Conference on Pattern Recognition,2004,4:384-387.
[41]Mikic.I,Cosman.P.C,Koqut.G.T,Trivedi.M.M.Moving shadow and object detection in traffic scenes[C].Proceedings of the 15th International Conference on Pattern Recognition,Barcelona,Spain,2000,1:321-324.
[42]Wei Zhiqiang,Ji Xiaopeng,Wang Peng.Real-time moving object detection for video monitoring systems[J].Journal of Systems Engineering and Electronics,2006,17(4):731-736.
[43]于成忠.视频序列中运动目标的提取与跟踪[D].中国优秀硕士学位论文全文数据库,东南大学,2006.
[44]R.Cucchiara,C.Grana,M.Piccardi,A.Prati.Detecting objects,shadows and ghost in video streams by exploiting color and motion information[C].In:Proceeding of the 11th International Conference on Image Analysis and Processing,2001:361-365.
[45]胡丹丹,高庆吉,支源.背景分割和阴影检测算法研究[J].中国图象图形学报,2008,13(8):1486-1490.
[46]刘鑫,刘辉,强振平,耿续涛.混合高斯模型和帧间差分相融的自适应背景模型[J].中国图象图形学报,2008,13(4):729-734.
[47]邵叶秦,任明武,杨静宇.基于多高斯分布的背景生成算法[J].计算机工程,2008,34(13):198-201.
[48]Nicolas Martel-Brisson,Andre Zaccarin.Moving cast shadow detection from a Gaussian Mixture Shadow Model[C].IEEE Computer Society Conference on Computer Vision and Pattern Recognition,2005:643-648.
[49]王典,程咏梅,杨涛,潘泉,赵春晖.基于混合高斯模型的运动阴影抑制算法[J].计算机应用,2006,26(5):1021-1024.
[50]Csaba Benedek,Tamas Sziranyi.Study on color space selection for detecting cast shadows in video surveillance[J].International Journal of Imaging Systems and Technology,2007,17(3):190-201.
[2]刘怀强.基于视频的车辆检测与跟踪技术研究[D].中国优秀硕士学位论文全文数库,中国海洋大学,2007.
[3]王素玉,沈兰荪.智能视觉监控技术研究进展[J].中国图象图形学报,2007,12(9):1505-1513.
[4]Haritaoglu I,Harwood D,Davis L.W4:Real-time surveillance of people and their activities[J].IEEE Trans Pattern Analysis and Machine Intelligence,2000,22(8):809-830.
[5]S.J.McKenna,S.Jabri,Z.Duric,A.Rosenfeld.Tracking groups of people[J].Computer Vision and Image Understanding,2000,80(1):42-56.
[6]Stauffer C,Grimson W.Adaptive background mixture models for real-time tracking[C].In:Proceeding of IEEE Conference on Computer Vision and Pattern Recognition,Fort Colline,Colorado,USA,1999:246-252.
[7]徐东彬,刘昌平,黄磊.基于概率统计自适应背景模型的运动目标检测方法[J].中国图象图形学报,2008,13(2):351-358.
[8]万峻甫,刘建伟,向怀坤,曹泉,王钧生.交通视频序列阴影检测算法研究[J].中国图象图形学报,2008,13(3):468-471.
[9]王成儒,顾广华.基于差分交集的视频对象分割与跟跟踪算法[J].光学技术,2004,30(5):564-566.
[10]倪峰.基于变化模板检测的视频对象分割算法研究[D].中国优秀硕士学位论文全文数据库,苏州大学,2007.
[11]石敏,易清明,刘金梅.一种基于边缘检测的去块效应算法[J].计算机工程与应用,2007,43(3):27-29.
[12]龚声蓉,刘纯平,王强.数字图像处理与分析[M].北京:清华大学出版社,2007.
[13]田宏阳,陈辉,马文静.一种动态场景下基于时空信息的视频对象提取算法[J].中 国图象图形学报,2007,12(9):1652-1657.
[14]杨文明.时空联合的视频对象分割[D].中国优秀硕士学位论文全文数数库,浙江大学,2006.
[15]高文,陈熙霖.计算机视觉--算法与系统原理[M].北京:清华大学出版社,1999.
[16]E P Ong.Robust optical flow computation based on least median of squares regression [J].Computer Vision,1999,31(1):51-82.
[17]邓玉春,姜昱明,张建荣.视频序列图像中运动对象分割综述[J].计算机应用研究,2005,1:8-11.
[18]Wollbom M,Mech R.Refined procedure for objective evaluation of video segmentation algorithm.Doc.ISO/IEC JTC 1/SC29/WG11 M3448,March,1998.
[19]J.L.Barron,D.J.Fleet,S.S.Beauchemin.Performance of optical techniques[J].International Journal of Computer Vision,1994,12(1):43-77.
[20]Simon Denman,Vinod Chandran,Sridha Sridharan.An adaptive optical flow technique for person tracking systems[J].Pattern Recognition Letters,2007,28(10):1232-1239.
[21]E.Jayabalan,Dr.A.Krishnan,R.Pugazendi.Non rigid object tracking in aerial videos by combined snaked and optical flow technique[J].Computer Graphics,Imaging and Visualization,2007,21(6):388-396.
[22]吴新根,罗立民.一种改进的光流场计算方法[J].电子学报,2000,28(1):130-132.
[23]Nils Papenberg.Highly accurate optic flow computation with theoretically justified warping[J].International Journal of Computer Vision,2006,67(2):141-158.
[24]杨国亮,王志良,牟世堂,解仑,刘冀伟.一种改进的光流算法[J].计算机工程,2006,32(15):187-189.
[25]曾浩,高秀娟,曾孝平.快速运动估计中一种改进的块匹配免疫算法[J].计算机应用,2008,28(8):2147-2149.
[26]洪波,余松煜.基于对象的菱形搜索运动估计方法[J].数据采集与处理,2001,16(1):110-114.
[27]Gong Sheng-rong,Zhou xiang.HDS:A fast and hierarchical diamond search algorithm in video motion estimation[C].Proceedings of SPIE - The International Society for Optical Engineering,v 6044,MIPPR 2005:Image Analysis Techniques,2005:60440D-1-60440D-8.
[28]李炜,周兵,李波.运动适量场自适应搜索算法[J].计算机学报,2003,23(2):168-173.
[29]章毓晋.图像工程(上册)图像处理[M].北京:清华大学出版社,2007.
[30]肖梅,韩崇昭,张雷.一种视频序列的背景提取算法[J].光电工程,2005,32(4):78-80.
[31]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,2007,12(9):1585-1589.
[32]张海青,李厚强.基于均匀混合模型的运动检测[J].计算机工程与应用,2007,43(9):28-30.
[33]Peng Suo,Yanjiang Wang.An improved adaptive background modeling algorithm based on Gaussian Mixture Model[C].In:Proceeding of the 9th International Conference on Signal Processing,2008:1436-1439.
[34]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.
[35]曾艳,于濂.一种新的道路交通背景提取算法及研究[J].中国图象图形学报,2008,13(3):593-599.
[36]杜友田,陈峰,徐文立.基于区域的运动阴影检测方法[J].清华大学学报,2006,46(1):141-144.
[37]万峻甫,刘建伟,向怀坤,曹泉,王钧生.交通视频序列阴影检测算法研究[J].中国图象图形学报,2008,13(3):467-471.
[38]Cavallaro A,Salvador E,Ebrahimi T.Shadow-aware object-based video processing[C].IEE Proceedings-Vision Image Signal Process,2005,152(4):398-406.
[39]陈柏生,陈锻生.基于归一化RGB彩色模型的运动阴影检测[J].计算机应用,2006,26(8):1879-1881.
[40]Siala.k,Chakchouk.M,Chaieb.F,Besbes.O.Moving shadows detection with support vector domain description in the color ratios space[C].In:Proceedings of the 17th International Conference on Pattern Recognition,2004,4:384-387.
[41]Mikic.I,Cosman.P.C,Koqut.G.T,Trivedi.M.M.Moving shadow and object detection in traffic scenes[C].Proceedings of the 15th International Conference on Pattern Recognition,Barcelona,Spain,2000,1:321-324.
[42]Wei Zhiqiang,Ji Xiaopeng,Wang Peng.Real-time moving object detection for video monitoring systems[J].Journal of Systems Engineering and Electronics,2006,17(4):731-736.
[43]于成忠.视频序列中运动目标的提取与跟踪[D].中国优秀硕士学位论文全文数据库,东南大学,2006.
[44]R.Cucchiara,C.Grana,M.Piccardi,A.Prati.Detecting objects,shadows and ghost in video streams by exploiting color and motion information[C].In:Proceeding of the 11th International Conference on Image Analysis and Processing,2001:361-365.
[45]胡丹丹,高庆吉,支源.背景分割和阴影检测算法研究[J].中国图象图形学报,2008,13(8):1486-1490.
[46]刘鑫,刘辉,强振平,耿续涛.混合高斯模型和帧间差分相融的自适应背景模型[J].中国图象图形学报,2008,13(4):729-734.
[47]邵叶秦,任明武,杨静宇.基于多高斯分布的背景生成算法[J].计算机工程,2008,34(13):198-201.
[48]Nicolas Martel-Brisson,Andre Zaccarin.Moving cast shadow detection from a Gaussian Mixture Shadow Model[C].IEEE Computer Society Conference on Computer Vision and Pattern Recognition,2005:643-648.
[49]王典,程咏梅,杨涛,潘泉,赵春晖.基于混合高斯模型的运动阴影抑制算法[J].计算机应用,2006,26(5):1021-1024.
[50]Csaba Benedek,Tamas Sziranyi.Study on color space selection for detecting cast shadows in video surveillance[J].International Journal of Imaging Systems and Technology,2007,17(3):190-201.