基于多特征融合与ROI预测的红外目标跟踪算法
|
摘要
针对当前红外目标检测与跟踪算法存在场景自适应能力弱、专用性强,以及在大视场条件下,首帧图像中小目标误检率高的问题,提出一种红外序列图像目标自适应阈值分割、检测与跟踪方法.选取目标移动速度、目标轮廓的面积和周长、以及自适应分割阈值与感兴趣区域位置为动态变量,建立动态决策准则.采用首帧目标检测算法计算出序列图像的第一帧图像目标的静态变量和部分动态变量,再采用改进的局部自适应阈值分割算法分割后续帧图像,然后利用静态与动态决策准则筛选出分割后的真实目标,最后计算并更新动态决策准则.红外靶标测试结果表明:该方法对不同场景具有较好的适应性,四个场景平均跟踪准确率为95.81%,微机平台平均每帧处理时间为10.93ms,嵌入式平台为26.79ms.
A method of adaptive threshold segmentation,detection and tracking for infrared sequence images is proposed to solve the problems that the current infrared target detection and tracking algorithm has weak scene adaptability,strong specificity and high false detection rate of small target in the first frame image under large field of view.The density,rectangularity and Hu invariant moments are selected as static variables to establish static decision criteria.The target moving speed,area and perimeter of target contour,adaptive segmentation threshold and location of ROI are selected as dynamic variables to establish dynamic decision criteria.The first frame target detection algorithm is used to calculate the target static variable and some of the dynamic features of the first frame image.The subsequent frame images are segmented by the improved local adaptive threshold segmentation algorithm and then the static and dynamic decision criteria are used to screen out the segmentation.Finally,the dynamic decision criteria are calculated and updated.The infrared target test results show that the method has good adaptability to different scenarios.By using this algorithm,the average tracking accuracy of the four scenarios is 95.81%,the average processing time per frame is 10.93 ms on microcomputer platform and26.79 ms on embedded platform respectively.
A method of adaptive threshold segmentation,detection and tracking for infrared sequence images is proposed to solve the problems that the current infrared target detection and tracking algorithm has weak scene adaptability,strong specificity and high false detection rate of small target in the first frame image under large field of view.The density,rectangularity and Hu invariant moments are selected as static variables to establish static decision criteria.The target moving speed,area and perimeter of target contour,adaptive segmentation threshold and location of ROI are selected as dynamic variables to establish dynamic decision criteria.The first frame target detection algorithm is used to calculate the target static variable and some of the dynamic features of the first frame image.The subsequent frame images are segmented by the improved local adaptive threshold segmentation algorithm and then the static and dynamic decision criteria are used to screen out the segmentation.Finally,the dynamic decision criteria are calculated and updated.The infrared target test results show that the method has good adaptability to different scenarios.By using this algorithm,the average tracking accuracy of the four scenarios is 95.81%,the average processing time per frame is 10.93 ms on microcomputer platform and26.79 ms on embedded platform respectively.
引文
[1]POJALA C,SOMNATH S.Neighborhood supported model level fuzzy aggregation for moving object segmentation[J].IEEE Transactions on Image Processing,2014,23(2):645-657.
[2]LI Wan-yi,WANG Peng,QIAO Hong.A survey of visual attention based methods for object tracking[J].Acta Automatica Sinica,2014,40(4):561-576.黎万义,王鹏,乔红.引入视觉注意机制的目标跟踪方法综述[J].自动化学报,2014,40(4):561-576.
[3]YIN Hong-peng,CHEN Bo,CHAI Yi,et al.Vision-based object detection and tracking:a review[J].Acta Automatica Sinica,2016,42(10):1466-1489.尹宏鹏,陈波,柴毅,等.基于视觉的目标检测与跟踪综述[J].自动化学报,2016,42(10):1466-1489.
[4]LI Cheng,LI Jian-xun,TONG Zhong-xiang,et al.Research on partial image recognition and tracking in infrared imaging terminal guidance[J].Acta Armamentarii,2015,36(7):1213-1221.李成,李建勋,童中翔,等.红外成像制导末端局部图像识别跟踪研究[J].兵工学报,2015,36(7):1213-1221.
[5]BOLME D S,BEVERRDGE J R,DRAPER B A,et al.Visual object tracking using adaptive correlation filters[C].IEEE Computer Society Conference on Computer Vision and Pattern Recognition,2010:2544-2550.
[6]HENRIQUES J F,CASERIO R,MARTINS P,et al.High-speed tracking with kernelized correlation filters[J].IEEETransactions on Pattern Analysis and Machine Intelligence,2015,37(3):583-596.
[7]XIONG Chang-zhen,ZHAO Lu-lu,GUO Fen-hong.Kernelized correlation filters tracking based on adaptive feature fusion[J].Journal of Computer-Aided Design&Computer Graphics,2017,29(6):1068-1074.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.
[8]LEE E,YOO H J,GU E,et al.Moving dim-target tracking algorithm using template matching[C].IEEE International Conference on Technological Advances in Electrical,2013:294-297.
[9]JIN Ze-fen-fen,HOU Zhi-qiang,YU Wang-sheng,et al.Multiple feature fusion based on covariance matrix for visual tracking[J].Acta Optica Sinca,2017,37(9):0915005.金泽芬芬,侯志强,余旺盛,等.基于协方差矩阵的多特征融合跟踪算法[J].光学学报,2017,37(9):0915005.
[10]ZHANG Shuai,LIU Bing-qi,HUANG Fu-yu,et al.Gradient entropy weighted centroid method for air target tracking in infrared wide field[J].Acta Photonica Sinica,2018,47(11):1110002.张帅,刘秉琦,黄富瑜,等.大视场红外空中小目标梯度熵权质心跟踪方法[J].光子学报,2018,47(11):1110002.
[11]LI Cheng-mei,BAI Hong-yang,GUO Hong-wei,et al.Moving object detection and tracking based on improved optical flow method[J].Chinese Journal of Scientific Instrument,2018,39(5):249-256.李成美,白宏阳,郭宏伟,等.一种改进光流法的运动目标检测及跟踪算法[J].仪器仪表学报,2018,39(5):249-256.
[12]MA C,HUANG J B,YANG X K,et al.Hierarchical convolutional features for visual tracking[C].IEEE International Conference on Computer Vision,2015:3074-3082.
[13]DANELLJAN M,ROBINSON A,KHAN F S,et al.Beyond correlation filters:Learning continuous convolution operators for visual tracking[C].14th European Conference on Computer Vision,2016:472-488.
[14]DANELLJAN M,BHAT G,KHAN F S,et al.ECO:Efficient convolution operators for tracking[C].IEEEConference on Computer Vision and Pattern Recognition,2017:6931-6939.
[15]ZHANG L,SUGANTHAN P N.Robust visual tracking via co-trained Kernelized correlation filters[J].Pattern Recognition,2017,69:82-93.
[16]LI H X,LI Y,PORIKLI F.Deep track:learning discriminative feature representations online for robust visual tracking[J].IEEE Transactions on Image Processing,2016,25(4):1834-1848.
[17]GUAN Hao,XUE Xiang-yang,AN Zhi-yong.Advances on application of deep learning for video object tracking[J].Acta Automatica Sinica,2016,42(6):834-847.管皓,薛向阳,安志勇.深度学习在视频目标跟踪中的应用进展与展望[J].自动化学报,2016,42(6):834-847.
[18]CHEN Fa-yang,XUE Hu.Design of low cost guidance and control system for precision guided munitions[J].Fire Control&Command Control,2017,42(11):178-180.陈法扬,薛琥.精确制导弹药低成本制导控制系统设计[J].火力与指挥控制,2017,42(11):178-180.
[19]JIA Jin-yan,CHEN Hai-feng,CHEN Ya-qing,et al.Integrated design of a new missile borne guidance information processing system[J].Winged Missiles Journal,2015,8:83-86.贾金艳,陈海峰,陈亚卿,等.新型弹载制导信息处理系统一体化设计[J].飞航导弹,2015,8:83-86.
[20]ZHANG Jun-yang,WANG Hui-li,GUO Yang,et al.Review of deep learning[J].Application Research of Computers,2018,35(7):1921-1936.张军阳,王慧丽,郭阳,等.深度学习相关研究综述[J].计算机应用研究,2018,35(7):1921-1936.
[21]MA Tian-yi,ZHANG Hui-xiang,SONG Min-min,et al.Anti-occluded infrared target tracking with salient feature space[J].Infrared and Laser Engineering,2017,46(3):0304002.马天义,张会香,宋敏敏,等.引入显著特征空间的抗遮挡红外目标跟踪[J].红外与激光工程,2017,46(3):0304002.
[22]YI Xiang,WANG Bing-jian.Fast infrared and dim target detection algorithm based on multi-feature[J].Acta Photonica Sinica,2017,46(6):0610002.易翔,王炳健.基于多特征的快速红外弱小目标检测算法[J].光子学报,2017,46(6):0610002.
[23]LEI Jian-feng,WANG Wei.Research and implementation of image threshold segmentation based on OpenCV[J].Modern Electronics Technique,2013,36(24):73-76.雷建锋,汪伟.基于OpenCV的图像阈值分割研究与实现[J].现代电子技术,2013,36(24):73-76.
[24]OTSU N.A threshold selection method from gary-level histograms[J].IEEE Transactions on System,Man,and Cybernetics,1979,9(1):62-66.
[25]ZHANG J,HU J L.Image segmentation based on 2D Otsu method with histogram analysis[C].IEEE International Conference on Computer Science and Software Engineering,2008,6:105-108.
[26]DO Y K,JONG W P.Connectivity-based local adaptive thresholding for carotid artery segmentation using MRA images[J].Image and Vision Computing,2005,9(5):1277-1287.
[2]LI Wan-yi,WANG Peng,QIAO Hong.A survey of visual attention based methods for object tracking[J].Acta Automatica Sinica,2014,40(4):561-576.黎万义,王鹏,乔红.引入视觉注意机制的目标跟踪方法综述[J].自动化学报,2014,40(4):561-576.
[3]YIN Hong-peng,CHEN Bo,CHAI Yi,et al.Vision-based object detection and tracking:a review[J].Acta Automatica Sinica,2016,42(10):1466-1489.尹宏鹏,陈波,柴毅,等.基于视觉的目标检测与跟踪综述[J].自动化学报,2016,42(10):1466-1489.
[4]LI Cheng,LI Jian-xun,TONG Zhong-xiang,et al.Research on partial image recognition and tracking in infrared imaging terminal guidance[J].Acta Armamentarii,2015,36(7):1213-1221.李成,李建勋,童中翔,等.红外成像制导末端局部图像识别跟踪研究[J].兵工学报,2015,36(7):1213-1221.
[5]BOLME D S,BEVERRDGE J R,DRAPER B A,et al.Visual object tracking using adaptive correlation filters[C].IEEE Computer Society Conference on Computer Vision and Pattern Recognition,2010:2544-2550.
[6]HENRIQUES J F,CASERIO R,MARTINS P,et al.High-speed tracking with kernelized correlation filters[J].IEEETransactions on Pattern Analysis and Machine Intelligence,2015,37(3):583-596.
[7]XIONG Chang-zhen,ZHAO Lu-lu,GUO Fen-hong.Kernelized correlation filters tracking based on adaptive feature fusion[J].Journal of Computer-Aided Design&Computer Graphics,2017,29(6):1068-1074.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.
[8]LEE E,YOO H J,GU E,et al.Moving dim-target tracking algorithm using template matching[C].IEEE International Conference on Technological Advances in Electrical,2013:294-297.
[9]JIN Ze-fen-fen,HOU Zhi-qiang,YU Wang-sheng,et al.Multiple feature fusion based on covariance matrix for visual tracking[J].Acta Optica Sinca,2017,37(9):0915005.金泽芬芬,侯志强,余旺盛,等.基于协方差矩阵的多特征融合跟踪算法[J].光学学报,2017,37(9):0915005.
[10]ZHANG Shuai,LIU Bing-qi,HUANG Fu-yu,et al.Gradient entropy weighted centroid method for air target tracking in infrared wide field[J].Acta Photonica Sinica,2018,47(11):1110002.张帅,刘秉琦,黄富瑜,等.大视场红外空中小目标梯度熵权质心跟踪方法[J].光子学报,2018,47(11):1110002.
[11]LI Cheng-mei,BAI Hong-yang,GUO Hong-wei,et al.Moving object detection and tracking based on improved optical flow method[J].Chinese Journal of Scientific Instrument,2018,39(5):249-256.李成美,白宏阳,郭宏伟,等.一种改进光流法的运动目标检测及跟踪算法[J].仪器仪表学报,2018,39(5):249-256.
[12]MA C,HUANG J B,YANG X K,et al.Hierarchical convolutional features for visual tracking[C].IEEE International Conference on Computer Vision,2015:3074-3082.
[13]DANELLJAN M,ROBINSON A,KHAN F S,et al.Beyond correlation filters:Learning continuous convolution operators for visual tracking[C].14th European Conference on Computer Vision,2016:472-488.
[14]DANELLJAN M,BHAT G,KHAN F S,et al.ECO:Efficient convolution operators for tracking[C].IEEEConference on Computer Vision and Pattern Recognition,2017:6931-6939.
[15]ZHANG L,SUGANTHAN P N.Robust visual tracking via co-trained Kernelized correlation filters[J].Pattern Recognition,2017,69:82-93.
[16]LI H X,LI Y,PORIKLI F.Deep track:learning discriminative feature representations online for robust visual tracking[J].IEEE Transactions on Image Processing,2016,25(4):1834-1848.
[17]GUAN Hao,XUE Xiang-yang,AN Zhi-yong.Advances on application of deep learning for video object tracking[J].Acta Automatica Sinica,2016,42(6):834-847.管皓,薛向阳,安志勇.深度学习在视频目标跟踪中的应用进展与展望[J].自动化学报,2016,42(6):834-847.
[18]CHEN Fa-yang,XUE Hu.Design of low cost guidance and control system for precision guided munitions[J].Fire Control&Command Control,2017,42(11):178-180.陈法扬,薛琥.精确制导弹药低成本制导控制系统设计[J].火力与指挥控制,2017,42(11):178-180.
[19]JIA Jin-yan,CHEN Hai-feng,CHEN Ya-qing,et al.Integrated design of a new missile borne guidance information processing system[J].Winged Missiles Journal,2015,8:83-86.贾金艳,陈海峰,陈亚卿,等.新型弹载制导信息处理系统一体化设计[J].飞航导弹,2015,8:83-86.
[20]ZHANG Jun-yang,WANG Hui-li,GUO Yang,et al.Review of deep learning[J].Application Research of Computers,2018,35(7):1921-1936.张军阳,王慧丽,郭阳,等.深度学习相关研究综述[J].计算机应用研究,2018,35(7):1921-1936.
[21]MA Tian-yi,ZHANG Hui-xiang,SONG Min-min,et al.Anti-occluded infrared target tracking with salient feature space[J].Infrared and Laser Engineering,2017,46(3):0304002.马天义,张会香,宋敏敏,等.引入显著特征空间的抗遮挡红外目标跟踪[J].红外与激光工程,2017,46(3):0304002.
[22]YI Xiang,WANG Bing-jian.Fast infrared and dim target detection algorithm based on multi-feature[J].Acta Photonica Sinica,2017,46(6):0610002.易翔,王炳健.基于多特征的快速红外弱小目标检测算法[J].光子学报,2017,46(6):0610002.
[23]LEI Jian-feng,WANG Wei.Research and implementation of image threshold segmentation based on OpenCV[J].Modern Electronics Technique,2013,36(24):73-76.雷建锋,汪伟.基于OpenCV的图像阈值分割研究与实现[J].现代电子技术,2013,36(24):73-76.
[24]OTSU N.A threshold selection method from gary-level histograms[J].IEEE Transactions on System,Man,and Cybernetics,1979,9(1):62-66.
[25]ZHANG J,HU J L.Image segmentation based on 2D Otsu method with histogram analysis[C].IEEE International Conference on Computer Science and Software Engineering,2008,6:105-108.
[26]DO Y K,JONG W P.Connectivity-based local adaptive thresholding for carotid artery segmentation using MRA images[J].Image and Vision Computing,2005,9(5):1277-1287.