一种融合卡尔曼滤波的改进时空上下文跟踪算法
收稿日期: 2016-04-11
修回日期: 2015-06-27
网络出版日期: 2016-06-28
基金资助
国家自然科学基金（11272256，61005062，60805034）
A tracking algorithm of improved spatio-temporal context with Kalman filter
Received date: 2016-04-11
Revised date: 2015-06-27
Online published: 2016-06-28
Supported by
National Natural Science Foundation of China (11272256, 61005062, 60805034)
针对时空上下文跟踪算法对快速运动和遭受严重遮挡目标的跟踪精度下降问题，提出一种融合卡尔曼滤波的改进时空上下文跟踪算法。首先人工标记目标所在的矩形区域，然后利用改进的时空上下文算法对目标进行稳定跟踪，在跟踪过程中，基于连续两帧图像灰度的欧氏距离判定目标跟踪状态，当判定目标遭受严重遮挡时，利用卡尔曼滤波进行预测估计。算法对噪声有一定的容忍度，通过降低噪声对跟踪过程的影响，能够得到更优的目标区域。仿真实验结果表明：本文算法适用于不同光照强度下高速、高机动目标的稳定跟踪，并且对目标的尺度变化和短时严重遮挡具有鲁棒性。算法帧平均耗时为34.07 ms；帧几何中心平均误差为5.43 pixel，比时空上下文算法减少70.2%；目标轮廓面积平均误差为13.08%，比时空上下文算法减少52.7%。
赵洲 , 黄攀峰 , 陈路 . 一种融合卡尔曼滤波的改进时空上下文跟踪算法[J]. 航空学报, 2017 , 38(2) : 320306 -320316 . DOI: 10.7527/S1000-6893.2016.0202
For the rapid target suffering from severe occlusion, the tracking accuracy of spatio-temporal context algorithm decreases. A novel tracking algorithm of improved spatio-temporal context with Kalman filter is proposed in the paper. The rectangular region of the tracking object is manually marked at the first frame, and the improved spatio-temporal context algorithm is then applied to track the target. The Euclidean distance of the image intensity in two consecutive frames determines the state of the target in the tracking process. We apply Kalman filter to reduce the influence of noise and predict and estimate the possible position of the target under severe occlusion, and obtain better rectangular region of the tracking object. The experimental results show that the algorithm of improved spatio-temporal context with Kalman filter can be used for high speed and highly maneuvering tracking target with different light intensities, and is robust for the target with varied scale and severe occlusion. Time consumption per frame is 34.07 ms. Geometric center error per frame is 5.43 pixel, 70.2% less than that via the spatio-temporal context algorithm. The contour area per frame is 13.08%, 52.7% less than that via the spatio-temporal context algorithm.
[1] ZHOU H Y, YUAN Y, ZHANG Y, et al. Non-rigid object tracking in complex scenes[J]. Pattern Recognition Letters, 2009, 30(2):98-102.
[2] SIVIC J, SCHAFFALITZKY F, ZISSERMAN A. Object level grouping for video shots[J]. International Journal of Computer Vision, 2006, 67(2):189-210.
[3] 张焕龙, 胡士强, 杨国胜. 基于外观模型学习的视频目标跟踪方法综述[J]. 计算机研究与发展, 2015, 52(1):177-190. ZHANG H L, HU S Q, YANG G S. Video object tracking based on appearance models learning[J]. Journal of Computer Research and Development, 2015, 52(1):177-190(in Chinese).
[4] ZIMMERMANN K, MATAS J, SVOBODA T. Tracking by an optimal sequence of linear predictors[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2009, 31(4):677-692.
[5] ELLIS L, DOWSON N, MATAS J, et al. Linear regression and adaptive appearance models for fast simultaneous modelling and tracking[J]. International Journal of Computer Vision, 2011, 95(2):154-179.
[6] KALAL Z, MIKOLAJCZYK K, MATAS J. Forward-backward error:Automatic detection of tracking failures[C]//Proceedings of 2010 the 20th International Conference on Pattern Recognition (ICPR). Piscataway, NJ:IEEE Press, 2010:2756-2759.
[7] ZHANG L, VAN DER MAATEN L. Preserving structure in model-free tracking[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2014, 36(4):756-769.
[8] TZIMIROPOULOS G, ZAFEIRIOU S, PANTIC M. Sparse representations of image gradient orientations for visual recognition and tracking[C]//Proceedings of 2011 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops (CVPRW). Piscataway, NJ:IEEE Press, 2011:26-33.
[9] LIWICKI S, ZAFEIRIOU S, TZIMIROPOULOS G, et al. Efficient online subspace learning with an indefinite kernel for visual tracking and recognition[J]. IEEE Transactions on Neural Networks and Learning Systems, 2012, 23(10):1624-1636.
[10] MEI X, LING H B. Robust visual tracking and vehicle classification via sparse representation[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2011, 33(11):2259-2272.
[11] HORBERT E, REMATAS K, LEIBE B. Level-set person segmentation and tracking with multi-region appearance models and top-down shape information[C]//Proceedings of 2011 IEEE International Conference on Computer Vision (ICCV). Piscataway, NJ:IEEE Press, 2011:1871-1878.
[12] NICOLAS P, AURELIE B. Tracking with occlusions via graph cuts[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2011, 33(1):144-157.
[13] POUYA B, SEYED A C, MUSA B M M. Upper body tracking using KLT and Kalman filter[J]. Procedia Computer Science, 2012, 13:185-191.
[14] WANG Y, LIU G. Head pose estimation based on head tracking and the Kalman filter[J]. Physics Procedia, 2011, 22:420-427.
[15] FU Z X, HAN Y. Centroid weighted Kalman filter for visual object tracking[J]. Measurement, 2012, 45(4):650-655.
[16] SU Y Y, ZHAO Q J, ZHAO L J, et al. Abrupt motion tracking using a visual saliency embedded particle filter[J]. Pattern Recognition, 2014, 47(5):1826-1834.
[17] 蔡佳, 黄攀峰. 基于改进SURF和P-KLT的特征点实时跟踪方法研究[J]. 航空学报, 2013, 34(5):1204-1214. CAI J, HUANG P F. Research of real-time feature point tracking method based on the combination of improved SURF and P-KLT algorithm[J]. Acta Aeronautica et Astronautica Sinica, 2013, 34(5):1204-1214(in Chinese).
[18] 高羽, 张建秋, 尹建君. 机动目标的多项式预测模型及其跟踪算法[J]. 航空学报, 2009, 30(8):1479-1489. GAO Y, ZHANG J Q, YIN J J. Polynomial prediction model and tracking algorithm of maneuver target[J]. Acta Aeronautica et Astronautica Sinica, 2009, 30(8):1479-1489(in Chinese).
[19] 甘明刚, 陈杰, 王亚楠, 等. 基于Mean Shift算法和NMI特征的目标跟踪算法研究[J]. 自动化学报, 2010, 36(9):1332-1336. GAN M G, CHEN J, WANG Y N, et al. A target tracking algorithm based on mean shift and normalized moment of inertia feature[J]. Acta Automatic Sinica, 2010, 36(9):1332-1336(in Chinese).
[20] THANG B D, NAM V, GERARD M. Context tracker:Exploring supporters and distracters in unconstrained environments[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Piscataway, NJ:IEEE Press, 2011:1177-1184.
[21] YANG M, WU Y, HUA G. Context-aware visual tracking[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence (PAMI), 2009, 31(7):1195-1209.
[22] GRABNER H, MATAS J, VAN G L, et al. Tracking the invisible:Learning where the object might be[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Piscataway, NJ:IEEE Press, 2010:1285-1292.
[23] ZHANG K H, ZHANG L, LIU Q S, et al. Fast visual tracking via dense spatio-temporal context learning[C]//European Conference on Computer Vision (ECCV). Zurich:Springer, 2014:127-141.
[24] CHEN S Y. Kalman filter for robot vision:A survey[J]. IEEE Transactions on Industrial Electronics, 2012, 59(11):4409-4420.
[25] 王向华, 覃征, 杨新宇, 等. 基于多次卡尔曼滤波的目标自适应跟踪算法与仿真分析[J]. 系统仿真学报, 2008, 20(23):6458-6465. WANG X H, QIN Z, YANG X Y, et al. Adaptive algorithm based on multi-Kalman filter for target tracking and simulation analyses[J]. Journal of System Simulation, 2008, 20(23):6458-6465(in Chinese).
[26] WANG L, ZHANG Y, FENG J. On the Euclidean distance of images[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence (PAMI), 2005, 27(8):1334-1339.
[27] WU Y, LIM J, YANG M H. Online object tracking:A benchmark[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Piscataway, NJ:IEEE Press, 2013:2411-2418.
[28] WU Y, LIM J, YANG M H. Object tracking benchmark[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2015, 37(9):1834-1848.
[29] ZHANG K, ZHANG L, YANG M H. Real-time compressive tracking[C]//European Conference on Computer Vision (ECCV). Florence:Springer, 2012:864-877.
[30] KALAL Z, MATAS J, MIKOLAJCZYK K. P-N learning:Bootstrapping linary classifiers by structural constraints[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Piscataway, NJ:IEEE Press, 2010:49-56.
/
| 〈 |
|
〉 |
CJA