一种改进的ViBe运动目标检测算法
发布时间：2018-11-04 21:19
【摘要】：为抑制传统ViBe算法在检测运动目标时产生的"鬼影",提高监控视频运动目标检测的准确性,提出了一种改进的ViBe运动目标检测算法。该算法采用连续相邻的多帧图像序列代替传统ViBe算法中第一帧图像,构建背景模型,从根源上解决传统ViBe算法在运动目标检测中存在的"鬼影"问题。利用Canny边缘检测算子和形态学运算相结合的方式准确完整的提取运动区域,降低算法的复杂度且减少运动区域的提取时间。提出一种背景模型更新策略判定条件,提高背景模型的质量,消除高频扰动对运动目标检测的影响,从而实时保证运动目标检测的准确性和鲁棒性。实验结果表明:经过4种算法的对比,改进的ViBe算法能够有效的抑制鬼影,且在高频扰动的情况下能较好的适应动态背景,显著提高运动目标检测的准确性。
[Abstract]:In order to suppress the "ghost" generated by the traditional ViBe algorithm in detecting moving targets and improve the accuracy of moving target detection in surveillance video, an improved ViBe moving target detection algorithm is proposed. In this algorithm, successive adjacent multi-frame image sequences are used to replace the first frame image in the traditional ViBe algorithm, and the background model is constructed to solve the "ghost" problem in moving target detection based on the traditional ViBe algorithm. Using the combination of Canny edge detection operator and morphological operation, the motion region is extracted accurately and completely, which reduces the complexity of the algorithm and reduces the extraction time of the motion region. A decision condition of background model updating strategy is proposed to improve the quality of background model and eliminate the influence of high frequency disturbance on moving target detection, so that the accuracy and robustness of moving target detection can be guaranteed in real time. The experimental results show that the improved ViBe algorithm can effectively suppress ghost images, adapt to the dynamic background and improve the accuracy of moving target detection.
【作者单位】： 山东理工大学交通与车辆工程学院;
【基金】：国家自然科学基金资助项目(51508315) 山东省自然科学基金资助项目(ZR2016EL19)
【分类号】：TP391.41;U495
本文编号：2311166
[Abstract]:In order to suppress the "ghost" generated by the traditional ViBe algorithm in detecting moving targets and improve the accuracy of moving target detection in surveillance video, an improved ViBe moving target detection algorithm is proposed. In this algorithm, successive adjacent multi-frame image sequences are used to replace the first frame image in the traditional ViBe algorithm, and the background model is constructed to solve the "ghost" problem in moving target detection based on the traditional ViBe algorithm. Using the combination of Canny edge detection operator and morphological operation, the motion region is extracted accurately and completely, which reduces the complexity of the algorithm and reduces the extraction time of the motion region. A decision condition of background model updating strategy is proposed to improve the quality of background model and eliminate the influence of high frequency disturbance on moving target detection, so that the accuracy and robustness of moving target detection can be guaranteed in real time. The experimental results show that the improved ViBe algorithm can effectively suppress ghost images, adapt to the dynamic background and improve the accuracy of moving target detection.
【作者单位】： 山东理工大学交通与车辆工程学院;
【基金】：国家自然科学基金资助项目(51508315) 山东省自然科学基金资助项目(ZR2016EL19)
【分类号】：TP391.41;U495
【相似文献】
相关期刊论文 前1条
1 陈亮;吴善刚;肖英杰;曾祥X;张浩;;基于视频目标检测的桥梁防船撞主动预警技术研究[J];水运工程;2012年06期
相关博士学位论文 前1条
1 臧风妮;智能视频监控中海面舰船目标检测算法研究[D];中国海洋大学;2014年
相关硕士学位论文 前4条
1 孙美婷;用于公路智能卡口的快速运动目标检测算法研究[D];哈尔滨理工大学;2016年
2 丁智;车载视频监控运动目标检测与跟踪算法研究[D];湖南大学;2016年
3 樊大帅;智能交通系统中的运动目标检测与识别[D];天津大学;2012年
4 胡婷;基于视频技术的车辆检测跟踪及分类算法研究与实现[D];长安大学;2015年
，本文编号：2311166
本文链接：https://www.wllwen.com/kejilunwen/ruanjiangongchenglunwen/2311166.html
