一种解决波动式干扰影响的序列图像运动目标检测方法

周建锋; 苏小红; 马培军

doi:10.3724/SP.J.1146.2009.00095

一种解决波动式干扰影响的序列图像运动目标检测方法

doi: 10.3724/SP.J.1146.2009.00095

计量
- 文章访问数: 3973
- HTML全文浏览量: 96
- PDF下载量: 1342
- 被引次数: 0
出版历程
- 收稿日期: 2009-01-19
- 修回日期: 2009-07-09
- 刊出日期: 2010-02-19

A Moving Targets Detection Approach to Remove the Fluctuant Interference in Video Sequences

摘要

摘要: 为解决复杂环境下的诸如枝叶摇摆、摄像机抖动等波动式干扰对运动目标检测的影响问题，该文提出基于视频窗口切分与分类的序列图像运动目标检测算法。首先将序列图像切分为rc大小的视频窗口，然后提取窗口内区域图像累积帧间差矩阵的简单统计特征，针对每一帧序列图像，将视频窗口进行分类，把它们划分为运动目标窗口和非运动目标窗口(包括静止背景窗口和波动式干扰窗口)，最后将运动目标窗口合并为运动目标。该方法的优点是无需已知背景模型和运动目标大小、形状等任何先验信息。实验表明该算法能在摄像机抖动以及枝叶干扰等复杂环境下快速有效的检测出运动目标。
- 运动目标检测; 波动式干扰; 视频窗口; 累积帧间差
Abstract: In order to remove the fluctuant interference such as swaying branches，stochastic shaking of camera in complex environments when moving targets are detected in video sequences, A new algorithm of moving targets detection based on video windows partition and classification is proposed. First of all, the image sequences are divided into numbers of video windows at size rc. Then the simple statistical feature is extracted from the matrix of accumulated frame differences in the window. According to the feature, the video windows is divided into two categories in each frame，moving object windows and non-moving object windows including static background windows and fluctuant interference windows. Finally all moving object windows are merged into the moving targets. The advantage of this method is that no knowledge about the background model and object size or shape is necessary. The results show that the algorithm can rapidly and validly detect the moving objects in complex environments with the fluctuant interference such as swaying branches and stochastic shaking of camera.

HTML全文

参考文献(1)

Elgammal A, Duraiswami R, and Harwood D, et al.. Background and foreground modeling using nonparametric kernel density estimation for visual surveillance[J]. Proceedings of the IEEE, 2002, 90(7): 1153-1163.[2]Chalidabhongse T H, Kim K, and Harwood D, et al.. A perturbation method for evaluating background subtraction algorithms[C]. Proceedings of the Joint IEEE International Workshop on Visual Surveillance and Performance Evaluation of Tracking and Surveillance, Nice, France, 2003, 10: 11-12.[3]Tian Ying-li and Hampapur A. Robust salient motion detection with complex background for real-time video surveillance[C]. Proceedings of IEEE Workshop on Motion and Video Computing, Breckenridge, CO, United states, 2005, (2): 30-35.[4]Wixson L. Detecting salient motion by accumulating directionally consistent flow[J].IEEE Transactions on Pattern Analysis and Machine Intelligence.2000, 22(8):774-780[5]Li B, Yuan B Z, and Miao Z J. Moving object detection in dynamic scenes using nonparametric local kernel histogram estimation[C]. IEEE International Conference on Multimedia and Expo, Hannover, Germany, 2008: 1461-1464.[6]Monnet A and Mittal A, et al.. Background and subtraction of dynamic scenes[C]. Proceedings of 9th IEEE International Conference on Computer Vision, Nice, France, 2003, 2: 1305-1312.[7]Zhang Jian-guo and Gong Shao-gang. People detection in low-resolution video with non-stationary background[J].Image and Vision Computing.2009, 27(4):437-443[8]Stauffer C and Grimson W. Adaptive background mixture models for real-time tracking[C]. Proceedings of IEEE Conference on Computer Vision and Pattern Recognition, Fort Collins, Colorado, USA, 1999: 245-252.[9]Zivkovic Z. Improved adaptive gaussian mixture model for background subtraction[C]. Proceedings of International Conference on Pattern Recognition, Cambridge, UK, 2004, 2: 28-31.[10]Poppe C, Martens G, Lambert P, and Van de Walle R. Mixture models based background subtraction for video surveillance applications[C]. Proceedings 12th International Conference on Computer Analysis of Images and Patterns, Vienna, Austria, 2007: 28-35.

施引文献

资源附件(0)

访问统计