高速移动图像丢帧误差的矫正方法研究
|
摘要
在研究运动图像跟踪过程中,目标大小的变化,对得到定位目标位置产生了非线性干扰。采用传统基于均值漂移算法的运动图像跟踪过程中,在遇到这种干扰时,由于无法自动调整跟踪窗口的大小,导致跟踪定位准确性较低,提出基于异常值的运动图像跟踪中丢帧误差消除方法,采用改进的粒子滤波算法,在目标相邻区域利用窗口随机抽样获取样本集,将各样本转换成和其相应的特征向量,构成稀疏基集。通过正交匹配追踪运算稀疏系数得到信心度,确定目标位置,实现目标运动图像的跟踪。依据运动图像产生的频率获取相邻两帧运动参数间差值的最大值,分析该最大值是否保持在一定的区间内,判断该帧是否异常,并采用差值法调整异常值。仿真实验,结果表明,所提方法可增强运动图像跟踪的实时性和准确性。
In the process of motion image tracking, the change of target size has a nonlinear disturbance to the location of the target. The traditional image tracking method based on mean shift algorithm is unable to automatically adjust the size of the tracking window when meeting this kind of interference, which leads to a lower track positioning accuracy. Based on the movement of outliers in the image tracking lost frame error elimination method, an improved particle filter algorithm, adjacent area in the target window random sampling is used to collect the sample set and convert each sample to its corresponding eigenvectors to form a sparse matrix set. Confidence degree is obtained by orthogonal matching pursuit operation sparse coefficients to determine the target location and track the target motion image. According to the motion frequency of acquiring the maximum difference between two adjacent frame motion parameters, whether the maximum stay within a certain range is analyzed, whether the frame is unusual is determined, and the difference method is adopted to adjust the outliers. The simulation results show that the proposed method can enhance the real-time and accuracy of the motion image tracking.
In the process of motion image tracking, the change of target size has a nonlinear disturbance to the location of the target. The traditional image tracking method based on mean shift algorithm is unable to automatically adjust the size of the tracking window when meeting this kind of interference, which leads to a lower track positioning accuracy. Based on the movement of outliers in the image tracking lost frame error elimination method, an improved particle filter algorithm, adjacent area in the target window random sampling is used to collect the sample set and convert each sample to its corresponding eigenvectors to form a sparse matrix set. Confidence degree is obtained by orthogonal matching pursuit operation sparse coefficients to determine the target location and track the target motion image. According to the motion frequency of acquiring the maximum difference between two adjacent frame motion parameters, whether the maximum stay within a certain range is analyzed, whether the frame is unusual is determined, and the difference method is adopted to adjust the outliers. The simulation results show that the proposed method can enhance the real-time and accuracy of the motion image tracking.
引文
[1]李腾达,赵晓炜.随机体育运动图像捕捉点虚拟重建中凌乱误差控制[J].科技通报,2014,2(59):31-34.Li T D,Zhao X H.Random sports image capture point virtual reconstruction messy error control[J].Science and technology,2014,2(59):31-34.
[2]朱雄泳,周杰,谭洪舟.一种消除LCD运动图像去模糊模型极点的方法[J].自动化学报,2012,38(5):759-768.Zhu X Y,Zhou J,Tan H Z.A movement against the LCD image to the method of fuzzy model poles[J].Journal of Automation,2012,38(5):759-768.
[3]宋艳杰,彭辉,李少腾.数字化电能计量中丢帧误码延时分析与事件告警判断研究[J].电测与仪表,2014,1(20):1-6.Song Y J,Peng H,Li S T.Digital frame in electric energy metering error delay analysis and event alarm judgment study[J].Electric Measurement and Instrument,2014,1(20):1-6.
[4]乔志伟.能消除CT图像角部误差的斜变滤波器长度选择方法[J].计算机工程与设计,2012,33(4):1486-1490.Qiao Z W.Oblique Angle error of CT images can eliminate variable filter length selection method[J].Computer Engineering and Design,2012,33(4):1486-1490.
[5]侯金俊.电学实验误差的图像法分析及消除[J].物理教学探讨:中学教学教研专辑,2012,30(7):57-59.Hou J J.Electrical experimental error method of image analysis and eliminate[J].Journal of Physics Teaching to Explore:Album of the Middle School Teaching Research,2012,30(7):57-59.
[6]于涛,胡炳樑,高晓惠,等.高光谱图像运动误差实时校正处理平台设计与实现[J].光谱学与光谱分析,2012,32(8):2275-2279.Yu T,Hu B L Gao X H,et al.Hyperspectral image motion error correction processing platform design and implementation of real-time[J].Spectroscopy and Spectral Analysis,2012,32(8):2275-2279.
[7]陈壮杰,宋占杰,高成志,等.导航雷达近海回波图像丢帧内插技术研究[J].海洋技术学报,2011,30(1):68-71.Chen Z J,Song Z J,Gao C Z,et al.Offshore navigation radar echo image lost frame interpolation technology research[J].Journal of Marine Technology,2011,30(1):68-71.
[8]齐志泉,宋野,王来生.基于在线学习的目标跟踪方法研究[J].计算机应用研究,2010,27(2):770-771.Qi Z Q,Song Y,Wang L S.Target tracking method based on online learning study[J].Computer Application Research,2010,27(2):770-771.
[9]刘金根,柳鑫,李志强.立体视频信号丢帧的误差分析模型[J].武汉理工大学学报:信息与管理工程版,2011,33(4):513-516.Liu J G,Liu X,Li Z Q,.The error analysis model of stereo video signal is lost frame[J].Journal of Wuhan University of Science and Technology,Information and Management Engineering,2011,33(4):513-516.
[10]牛芗洁,黄永春.微弱运动目标的检测与跟踪识别算法研究[J].计算机仿真,2010,27(4):245-247.Niu X J,Huang Y C.Weak moving target detection and tracking of the identification algorithm study[J].Computer Simulation,2010,27(4):245-247.
[11]辛秀,方泽凯,杨芷华.基于MSK与MIMO技术的高速移动通信图像传输系统性能研究[J].科技视界,2015,13(3):83-84.Xin X,Fang Z K,Yang Z H.Based on MSK with MIMO technology of high-speed mobile communication image transmission system performance study[J].Science and Technology Horizon.2015,13(3):83-84.
[12]孙世洁,赵怀慈,郝明国.告诉移动无线图像循传输链路误差分析[J].电视技术,2013,11(37):166-169.Sun S J,Zhao H C,Hao M G.Tell mobile wireless image transmission path and link error analysis[J].Journal of TV Technology.2013,11(37):166-169.
[13]孙俊,何发智,陈晓,等.多特征提取逐步求精的高速移动目标跟踪算法[J].计算机辅助设计与图形学学报,2014,10(26):1747-1752.Sun J,He F Z,Chen X,et al.Many feature extraction incremental refinement of high speed moving target tracking algorithm[J].Journal of Computer-Aided Design and Graphics,2014,10(26):1747-1752.
[14]宋艳杰,彭辉,李少腾.数字化电能计量中丢帧误码延时分析与时间告警判断研究[J].电测与仪表,2014,20(51):1-6.Song Y J,Peng H,Li S T.Digital frame in electric energy metering error analysis and time delay alarm judgment study[J].Electric Logging and Instrumentation,2014,20(51):1-6.
[15]李莹.一种基于测量误差补偿矫正的为人机远程测距算法[J].计算机测量与控制,2012,9(9):2360-2362.Li Y.Based on the correct measurement error compensation for man-machine remote distance algorithm[J].Computer measurement and control,2012,9(9):2360-2362.
[16]何慧钧,张文强,邱晓欣,等.高速移动背景下目标零件的实时识别系统[J]计算机应用与软件,2012,11(29):153-155.He H J,Zhang W Q,Qiu X X,et al.Under the background of high-speed mobile target parts of real-time identification system[J].Computer Applications and Software,2012,11(29):153-155.
[17]谢宏亮,满君丰.基于motion的移动图像监控系统的实际与实现[J]通讯世界,2015,3(8):218-219.Xie H L,Man J F.The mobile video surveillance system based on motion of the real world and implementation[J].Communication,2015,3(8):218-219.
[18]宋勇,宋子房,周晓华.公路工程移动图像监控信息系统研究[J].软件,2014,1(35):79-82.Song Y,Song Z F,Zhou X H.Highway engineering mobile image monitoring information system software research[J].Computer Engineering&Software,2014,1(35):79-82.
[2]朱雄泳,周杰,谭洪舟.一种消除LCD运动图像去模糊模型极点的方法[J].自动化学报,2012,38(5):759-768.Zhu X Y,Zhou J,Tan H Z.A movement against the LCD image to the method of fuzzy model poles[J].Journal of Automation,2012,38(5):759-768.
[3]宋艳杰,彭辉,李少腾.数字化电能计量中丢帧误码延时分析与事件告警判断研究[J].电测与仪表,2014,1(20):1-6.Song Y J,Peng H,Li S T.Digital frame in electric energy metering error delay analysis and event alarm judgment study[J].Electric Measurement and Instrument,2014,1(20):1-6.
[4]乔志伟.能消除CT图像角部误差的斜变滤波器长度选择方法[J].计算机工程与设计,2012,33(4):1486-1490.Qiao Z W.Oblique Angle error of CT images can eliminate variable filter length selection method[J].Computer Engineering and Design,2012,33(4):1486-1490.
[5]侯金俊.电学实验误差的图像法分析及消除[J].物理教学探讨:中学教学教研专辑,2012,30(7):57-59.Hou J J.Electrical experimental error method of image analysis and eliminate[J].Journal of Physics Teaching to Explore:Album of the Middle School Teaching Research,2012,30(7):57-59.
[6]于涛,胡炳樑,高晓惠,等.高光谱图像运动误差实时校正处理平台设计与实现[J].光谱学与光谱分析,2012,32(8):2275-2279.Yu T,Hu B L Gao X H,et al.Hyperspectral image motion error correction processing platform design and implementation of real-time[J].Spectroscopy and Spectral Analysis,2012,32(8):2275-2279.
[7]陈壮杰,宋占杰,高成志,等.导航雷达近海回波图像丢帧内插技术研究[J].海洋技术学报,2011,30(1):68-71.Chen Z J,Song Z J,Gao C Z,et al.Offshore navigation radar echo image lost frame interpolation technology research[J].Journal of Marine Technology,2011,30(1):68-71.
[8]齐志泉,宋野,王来生.基于在线学习的目标跟踪方法研究[J].计算机应用研究,2010,27(2):770-771.Qi Z Q,Song Y,Wang L S.Target tracking method based on online learning study[J].Computer Application Research,2010,27(2):770-771.
[9]刘金根,柳鑫,李志强.立体视频信号丢帧的误差分析模型[J].武汉理工大学学报:信息与管理工程版,2011,33(4):513-516.Liu J G,Liu X,Li Z Q,.The error analysis model of stereo video signal is lost frame[J].Journal of Wuhan University of Science and Technology,Information and Management Engineering,2011,33(4):513-516.
[10]牛芗洁,黄永春.微弱运动目标的检测与跟踪识别算法研究[J].计算机仿真,2010,27(4):245-247.Niu X J,Huang Y C.Weak moving target detection and tracking of the identification algorithm study[J].Computer Simulation,2010,27(4):245-247.
[11]辛秀,方泽凯,杨芷华.基于MSK与MIMO技术的高速移动通信图像传输系统性能研究[J].科技视界,2015,13(3):83-84.Xin X,Fang Z K,Yang Z H.Based on MSK with MIMO technology of high-speed mobile communication image transmission system performance study[J].Science and Technology Horizon.2015,13(3):83-84.
[12]孙世洁,赵怀慈,郝明国.告诉移动无线图像循传输链路误差分析[J].电视技术,2013,11(37):166-169.Sun S J,Zhao H C,Hao M G.Tell mobile wireless image transmission path and link error analysis[J].Journal of TV Technology.2013,11(37):166-169.
[13]孙俊,何发智,陈晓,等.多特征提取逐步求精的高速移动目标跟踪算法[J].计算机辅助设计与图形学学报,2014,10(26):1747-1752.Sun J,He F Z,Chen X,et al.Many feature extraction incremental refinement of high speed moving target tracking algorithm[J].Journal of Computer-Aided Design and Graphics,2014,10(26):1747-1752.
[14]宋艳杰,彭辉,李少腾.数字化电能计量中丢帧误码延时分析与时间告警判断研究[J].电测与仪表,2014,20(51):1-6.Song Y J,Peng H,Li S T.Digital frame in electric energy metering error analysis and time delay alarm judgment study[J].Electric Logging and Instrumentation,2014,20(51):1-6.
[15]李莹.一种基于测量误差补偿矫正的为人机远程测距算法[J].计算机测量与控制,2012,9(9):2360-2362.Li Y.Based on the correct measurement error compensation for man-machine remote distance algorithm[J].Computer measurement and control,2012,9(9):2360-2362.
[16]何慧钧,张文强,邱晓欣,等.高速移动背景下目标零件的实时识别系统[J]计算机应用与软件,2012,11(29):153-155.He H J,Zhang W Q,Qiu X X,et al.Under the background of high-speed mobile target parts of real-time identification system[J].Computer Applications and Software,2012,11(29):153-155.
[17]谢宏亮,满君丰.基于motion的移动图像监控系统的实际与实现[J]通讯世界,2015,3(8):218-219.Xie H L,Man J F.The mobile video surveillance system based on motion of the real world and implementation[J].Communication,2015,3(8):218-219.
[18]宋勇,宋子房,周晓华.公路工程移动图像监控信息系统研究[J].软件,2014,1(35):79-82.Song Y,Song Z F,Zhou X H.Highway engineering mobile image monitoring information system software research[J].Computer Engineering&Software,2014,1(35):79-82.