基于核相关滤波算法的青鳉游泳跟踪研究
|
摘要
为获取鱼类游泳行为数据,采用增强的核相关滤波跟踪算法,对青鳉Oryzias latipes游泳轨迹进行跟踪研究,根据当前和之前帧信息训练出多核相关滤波器,然后与新输入帧进行相关性计算,得到的响应值就是预测的跟踪结果,响应值最大处即为跟踪目标位置,用该方法提取出3万帧图像中青鳉游泳位置,绘制青鳉游泳轨迹热力图,并分析不同规格青鳉的游泳领地。结果表明:增强的核相关滤波跟踪算法能够实时稳定地跟踪青鳉游泳行为,提取的青鳉游泳轨迹与实时记录轨迹高度吻合;试验过程中测试2万帧视频序列,806帧丢失,丢失率4.03%。本研究实现了对青鳉游泳位置定位跟踪,及时显示青鳉的游泳位置、速度和轨迹,对研究青鳉在特定环境下的生活习性提供了数据支持。
An enhanced kernelized correlation filter tracking algorithm is used to track the swimming trajectory of mekada Oryzias latipes to understand fish swimming behavior. Firstly, the multi-core correlation filter is established according to the current and previous frame information, and then the correlation response value is calculated by the new input frame which is the predicted tracking result, and the maximal response value is the tracking target position. The method realizes the real-time swimming location positioning and tracking of single and multiple medaka and also displays the swimming speed and track of the medaka in every frame of the picture. It also induces the location and speed data into the database for further data analysis, so as to produce a complete swimming speed and tracking curve of the medaka. The swimming position of mekada can be extracted from 30 000 frames, and the swimming trajectory thermodynamic diagram of mekada is drawn, with analysis of swimming areas of mekada with various sizes. The test result shows that the enhance kernelized correlation filter tracking algorithm is able to track the swimming behavior of medaka in a real-time and stable manner and the produced swimming track of the medaka is extremely similar to the real-time recorded track of the medaka through applying acquired data. In the test, 20 000 frames video sequences were measured, with 806 frames lost and the loss rate of 4.03%.
An enhanced kernelized correlation filter tracking algorithm is used to track the swimming trajectory of mekada Oryzias latipes to understand fish swimming behavior. Firstly, the multi-core correlation filter is established according to the current and previous frame information, and then the correlation response value is calculated by the new input frame which is the predicted tracking result, and the maximal response value is the tracking target position. The method realizes the real-time swimming location positioning and tracking of single and multiple medaka and also displays the swimming speed and track of the medaka in every frame of the picture. It also induces the location and speed data into the database for further data analysis, so as to produce a complete swimming speed and tracking curve of the medaka. The swimming position of mekada can be extracted from 30 000 frames, and the swimming trajectory thermodynamic diagram of mekada is drawn, with analysis of swimming areas of mekada with various sizes. The test result shows that the enhance kernelized correlation filter tracking algorithm is able to track the swimming behavior of medaka in a real-time and stable manner and the produced swimming track of the medaka is extremely similar to the real-time recorded track of the medaka through applying acquired data. In the test, 20 000 frames video sequences were measured, with 806 frames lost and the loss rate of 4.03%.
引文
[1] 冯春雷,李志国,黄洪亮,等.鱼类行为研究在捕捞中的应用[J].大连水产学院学报,2009,24(2):166-170.
[2] 于晓明,张秀梅.鱼类游泳能力测定方法的研究进展[J].南方水产科学,2011,7(4):76-84.
[3] 范良忠,刘鹰,余心杰,等.基于计算机视觉技术的运动鱼检测算法[J].农业工程学报,2011,27(7):226-230.
[4] 张胜茂,张衡,唐峰华,等.计算机视觉技术在监测鱼类游泳行为中的研究进展[J].大连海洋大学学报,2017,32(4):493-500.
[5] 郭静,罗华,张涛.机器视觉与应用[J].电子科技,2014,27(7):185-188.
[6] Kalman R E.A new approach to linear filtering and prediction problems[J].Journal of Basic Engineering,1960,82(1):35-45.
[7] Wang Naiyan,Shi Jianping,Yeung D Y,et al.Understanding and diagnosing visual tracking systems[C]//Proceedings of 2015 IEEE International Conference on Computer Vision.Santiago,Chile:IEEE,2015:3101-3109.
[8] Comaniciu D,Meer P.Meanshift:a robust approach toward feature space analysis[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2002,24(5):603-619.
[9] Johnson A E,Hebert M.Using spin images for efficient object recognition in cluttered 3D scenes[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,1999,21(5):433-449.
[10] Avidan S.Support vector tracking[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2004,26(8):1064-1072.
[11] 徐盼麟,韩军,童剑锋.基于单摄像机视频的鱼类三维自动跟踪方法初探[J].水产学报,2012,36(4):623-628.
[12] 黄一凡,陈欣,袁飞.鱼类应激行为作用下的水质视频监测分析系统[J].厦门大学学报:自然科学版,2017,56(4):584-589.
[13] 颜鹏东,谭均军,高柱,等.基于视频跟踪的竖缝式鱼道内鱼类运动行为分析[J].水生生物学报,2018,42(2):250-254.
[14] Henriques J F,Caseiro R,Martins P,et al.High-speed tracking with kernelized correlation filters[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2015,37(3):583-596.
[15] 李治国,路娜,郭晓芝,等.基于青鳉鱼行为的在线生物监测系统及综合预警平台建设研究[J].科学技术与工程,2017,17(7):288-292,298.
[16] 张融,郑宏远,李录,等.一种计量青鳉鱼胸鳍和尾鳍摆动频率和幅值的计算机视觉算法[J].生态毒理学报,2015,10(4):154-161.
[17] 张胜茂,赵申,张衡,等.基于视频分析的鱼类运动目标提取[J].渔业信息与战略,2017,32(1):44-50.
[18] 伊俊超,刘直芳.基于OpenCV的运动目标检测与跟踪[J].计算机工程与设计,2011,32(8):2817-2820.
[19] Nixon M S,Aguado A S.Feature extraction and image processing[M].Amsterdam:Academic Press,2008.
[20] Haddad R A,Akansu A N.A class of fast Gaussian binomial filters for speech and image processing[J].IEEE Transactions on Signal Processing,1991,39(3):723-727.
[21] Dalal N,Triggs B.Histograms of oriented gradients for human detection[C]//Proceedings of 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition.San Diego,CA,USA:IEEE,2005.
[22] Wu Yi,Lim J,Yang M H.Online object tracking:a benchmark[C]//Proceedings of 2013 IEEE Conference on Computer Vision and Pattern Recognition.Portland,OR,USA:IEEE,2013.
[23] 汪廷华,陈峻婷.核函数的选择研究综述[J].计算机工程与设计,2012,33(3):1181-1186.
[24] 禹文豪,艾廷华,杨敏,等.利用核密度与空间自相关进行城市设施兴趣点分布热点探测[J].武汉大学学报:信息科学版,2016,41(2):221-227.
[25] 董浩洋,张东戈,万贻平,等.战场态势热力图构建方法研究[J].指挥控制与仿真,2017,39(5):1-8.
[26] 邓超,宋金伟,孙瑞志,等.基于热力图的卷烟市场数据可视分析系统[J].烟草科技,2016,49(12):91-97.
[27] 邢运龙,李艾华,崔智高,等.改进核相关滤波的运动目标跟踪算法[J].红外与激光工程,2016,45(S1):214-221.
[28] 熊昌镇,赵璐璐,郭芬红.自适应特征融合的核相关滤波跟踪算法[J].计算机辅助设计与图形学学报,2017,29(6):1068-1074.
[2] 于晓明,张秀梅.鱼类游泳能力测定方法的研究进展[J].南方水产科学,2011,7(4):76-84.
[3] 范良忠,刘鹰,余心杰,等.基于计算机视觉技术的运动鱼检测算法[J].农业工程学报,2011,27(7):226-230.
[4] 张胜茂,张衡,唐峰华,等.计算机视觉技术在监测鱼类游泳行为中的研究进展[J].大连海洋大学学报,2017,32(4):493-500.
[5] 郭静,罗华,张涛.机器视觉与应用[J].电子科技,2014,27(7):185-188.
[6] Kalman R E.A new approach to linear filtering and prediction problems[J].Journal of Basic Engineering,1960,82(1):35-45.
[7] Wang Naiyan,Shi Jianping,Yeung D Y,et al.Understanding and diagnosing visual tracking systems[C]//Proceedings of 2015 IEEE International Conference on Computer Vision.Santiago,Chile:IEEE,2015:3101-3109.
[8] Comaniciu D,Meer P.Meanshift:a robust approach toward feature space analysis[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2002,24(5):603-619.
[9] Johnson A E,Hebert M.Using spin images for efficient object recognition in cluttered 3D scenes[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,1999,21(5):433-449.
[10] Avidan S.Support vector tracking[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2004,26(8):1064-1072.
[11] 徐盼麟,韩军,童剑锋.基于单摄像机视频的鱼类三维自动跟踪方法初探[J].水产学报,2012,36(4):623-628.
[12] 黄一凡,陈欣,袁飞.鱼类应激行为作用下的水质视频监测分析系统[J].厦门大学学报:自然科学版,2017,56(4):584-589.
[13] 颜鹏东,谭均军,高柱,等.基于视频跟踪的竖缝式鱼道内鱼类运动行为分析[J].水生生物学报,2018,42(2):250-254.
[14] Henriques J F,Caseiro R,Martins P,et al.High-speed tracking with kernelized correlation filters[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2015,37(3):583-596.
[15] 李治国,路娜,郭晓芝,等.基于青鳉鱼行为的在线生物监测系统及综合预警平台建设研究[J].科学技术与工程,2017,17(7):288-292,298.
[16] 张融,郑宏远,李录,等.一种计量青鳉鱼胸鳍和尾鳍摆动频率和幅值的计算机视觉算法[J].生态毒理学报,2015,10(4):154-161.
[17] 张胜茂,赵申,张衡,等.基于视频分析的鱼类运动目标提取[J].渔业信息与战略,2017,32(1):44-50.
[18] 伊俊超,刘直芳.基于OpenCV的运动目标检测与跟踪[J].计算机工程与设计,2011,32(8):2817-2820.
[19] Nixon M S,Aguado A S.Feature extraction and image processing[M].Amsterdam:Academic Press,2008.
[20] Haddad R A,Akansu A N.A class of fast Gaussian binomial filters for speech and image processing[J].IEEE Transactions on Signal Processing,1991,39(3):723-727.
[21] Dalal N,Triggs B.Histograms of oriented gradients for human detection[C]//Proceedings of 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition.San Diego,CA,USA:IEEE,2005.
[22] Wu Yi,Lim J,Yang M H.Online object tracking:a benchmark[C]//Proceedings of 2013 IEEE Conference on Computer Vision and Pattern Recognition.Portland,OR,USA:IEEE,2013.
[23] 汪廷华,陈峻婷.核函数的选择研究综述[J].计算机工程与设计,2012,33(3):1181-1186.
[24] 禹文豪,艾廷华,杨敏,等.利用核密度与空间自相关进行城市设施兴趣点分布热点探测[J].武汉大学学报:信息科学版,2016,41(2):221-227.
[25] 董浩洋,张东戈,万贻平,等.战场态势热力图构建方法研究[J].指挥控制与仿真,2017,39(5):1-8.
[26] 邓超,宋金伟,孙瑞志,等.基于热力图的卷烟市场数据可视分析系统[J].烟草科技,2016,49(12):91-97.
[27] 邢运龙,李艾华,崔智高,等.改进核相关滤波的运动目标跟踪算法[J].红外与激光工程,2016,45(S1):214-221.
[28] 熊昌镇,赵璐璐,郭芬红.自适应特征融合的核相关滤波跟踪算法[J].计算机辅助设计与图形学学报,2017,29(6):1068-1074.