基于多项式展开及Harris角点的轨迹跟踪系统

2017-08-30 02:54杨广学胡嘉维康守强梁亚琦王磊
哈尔滨理工大学学报 2017年3期

杨广学+胡嘉维+康守强+梁亚琦+王磊

摘 要:提出了一种基于多项式展开帧间估计和Harris角点检测的运动轨迹跟踪系统。系统先将视频监控图像输入系统,转化为灰度图,以便进行特征提取;接下来确定兴趣区域抓取目标点,提取ROI内角点坐标;最后确定角点的新坐标,定位目标在下一帧的位置,从而实现对目标的跟踪。通过对常用的光流检测算法和角点判断方式进行分析比较,最终选择多项式展开帧间估计和Harris角点检测算法作为图像配准和角点跟踪的方案,并进行参数优化,使系统的鲁棒性和得到了提升,并更能够适应复杂的环境条件。

关键词:多项式展开;角点;运动轨迹跟踪

DOI:10.15938/j.jhust.2017.03.009

中圖分类号: TP391.4

文献标志码: A

文章编号: 1007-2683(2017)03-0048-06

Abstract:A motion tracking system based on polynomial expansion inter-frame estimation and Harris corner detection is proposed. The system transforms the video monitoring image into grayscale in order to feature extract, then determines the interested area to grab the target and analyses the corner point and extract the ROI interior point coordinates, and finally confirms the new corner points by polynomial expansion to locate the position of the target in next frame. By means of analyzing and comparing common methods of light flow and corner point judgment, we select polynomial expansion frame and Harris corner detection as the image registration and the corner point tracking scheme, and conduct the parameter optimization. Through the new algorithm the system robustness and adaptability are promoted, and more able to adapt to complex environmental conditions.

Keywords:polynomial expansion; Harris corner; motion tracking

5 结 论

本文通过多项式展开的帧间估计算法和Harris角点算法,实现了运动轨迹实时跟踪系统,并使精度和适应性得到了提高。同时结合Python和OpenCV,使较为复杂的工程项目的实现难度降低。实验结果表明,该轨迹跟踪系统能够对运动物体进行有效跟踪。在后面的研究中,系统还可以加入多物体识别模块,使得该系统有更普遍的适用性。

参 考 文 献:

[1] 周志艳, 闫梦璐, 陈盛德, 等. Harris 角点自适应检测的水稻低空遥感图像配准与拼接算法[J]. 农业工程学报, 2015, 31(14): 186-193.

[2] 池利阳, 李民政, 吴新强. 一种运动目标检测实时性和鲁棒性的方法[J]. 计算机应用与软件, 2015, 32(2): 132-134.

[3] BERTHOLD K P. Horn, BRIAN G. Schunck. Determining Optical Flow. Artifical Intelligence, 1981(17):185-203.

[4] SHI J, TOMASI C. Good Features to Track, in: Computer Vision and Pattern Recognition, Seattle,1994:593-600.

[5] YILMAZ A, JAVED O, SHAH M. Object Tracking: A Survey [J]. Acm Computing Surveys, 2006, 38(4): 1-45 .

[6] SCHMID C, MOHR R, BACUHKHAGE C. Evaluation of Interest Point Detectors [J]. Iternational Journal of Computer Vision, 2000, 37(2): 151-172.

[7] 赵文哲, 秦世引. 视频运动目标检测方法的对比分析[J]. 科技导报, 2009 (10): 64-70.

[8] SUN D, ROTH S, BLACK M J. A Quantitative Analysis of Current Practices in Optical Flow Estimation and the Principles Behind them[J]. International Journal of Computer Vision, 2014, 106(2): 115-137.

[9] 诸亮, 周林强, 李一啸, 等. 一种基于低秩矩阵逼近的图像去噪算法[J]. 杭州电子科技大学学报, 2015, 35(4): 36-40.

[10]FARNEBACK G. Twoframe Motion Estimation Based on Polynomial Expansion[C]// In Scandinavian Conference on Image Analysis, 2003:363-370.

[11]王远军. 基于多项式展开的弥散张量图像配准 [D]. 上海:复旦大学, 2010:35-36.

[12]密聪杰, 刘洋, 薛向阳. 基于多帧图像的视频文字跟踪和分割算法[J]. 计算机研究与发展, 2006,43(9):1523-1529.

[13]谢勤岚, 曹汇敏. 结合两步估算和运动补偿的子像素运动估计[J]. 计算机工程与應用, 2009, 45(7): 190-191.

[14]FARNEBACK G: Polynomial Expansion for Orientation and Motion Estimation[D].Linoping: Linopings University,2002:41-56.

[15]BARRON J L, FLEET D J, BEAUCHEMIN S S. Performance of Optical Flow Techniques[C]// IEEE Computer Society Conference on Computer Vision & Pattern Recognition, Cvpr. 1994:236-242.

[16]MORAVEC H P. Towards Automatic Visual Obstacle Avoidance [C]// Cambridge: Proceedings of International Joint Conference on Artificial Intelligence,1977: 584-590.

[17]田原,梁德群,吴更石.直接基于灰度图像的多尺度角点检测方法[J].信号处理,2006,14(12):6-10.

[18]JOHN S N, MATHEW A A. Rooftop Extraction in Aerial Imagery Using Corners by Harris Corner Detector and SUSAN Corner Detector[J]. 2014,22(2):4395-4401.

[19]陈凯歌, 韩晓军, 张宇, 等. 基于 OpenCV 的 Harris 角点检测[J]. 河南工程学院学报: 自然科学版, 2011, 23(4): 53-56.

(编辑:温泽宇)