基于背景减除法的视频序列运动目标检测
发布时间：2018-11-19 13:27
【摘要】：视频序列中运动目标的检测是目标识别、标记和追踪的重要组成部分,背景减除法是运动目标检测中被广泛应用的算法。针对光线变化、噪声和局部运动等影响运动目标检测效果的问题,提出一种基于背景减除法的视频序列运动目标检测算法。该算法结合背景减除法和帧间差分法,对当前帧像素点的运动状态进行判断,分别对静止和运动的像素点进行替换和更新,采用最大类间方差(Otsu)法对差分图像进行目标提取,并使用数学形态学运算去除目标中的噪声和冗余信息。实验结果表明,所提算法对于视频序列中运动目标的检测具有较好的视觉效果和较高的准确度,能够克服局部运动以及噪声等缺陷。
[Abstract]:Moving target detection in video sequences is an important part of target recognition, marking and tracking. Background subtraction is a widely used algorithm in moving target detection. A video sequence moving target detection algorithm based on background subdivision is proposed to solve the problems such as the change of light noise and local motion. The algorithm combines background subtraction method and inter-frame difference method to judge the motion state of the current frame pixels and to replace and update the still and moving pixels respectively. The maximum inter-class variance (Otsu) method is used to extract the target from the differential image, and the mathematical morphological operation is used to remove the noise and redundant information from the target. Experimental results show that the proposed algorithm has better visual effect and higher accuracy for moving target detection in video sequences, and can overcome the defects of local motion and noise.
【作者单位】： 兰州理工大学电气工程与信息工程学院;西北民族大学数学与计算机科学学院;
【基金】：国家自然科学基金资助项目(61561042) 西北民族大学引进人才基金资助项目~~
【分类号】：TP391.41
本文编号：2342432
[Abstract]:Moving target detection in video sequences is an important part of target recognition, marking and tracking. Background subtraction is a widely used algorithm in moving target detection. A video sequence moving target detection algorithm based on background subdivision is proposed to solve the problems such as the change of light noise and local motion. The algorithm combines background subtraction method and inter-frame difference method to judge the motion state of the current frame pixels and to replace and update the still and moving pixels respectively. The maximum inter-class variance (Otsu) method is used to extract the target from the differential image, and the mathematical morphological operation is used to remove the noise and redundant information from the target. Experimental results show that the proposed algorithm has better visual effect and higher accuracy for moving target detection in video sequences, and can overcome the defects of local motion and noise.
【作者单位】： 兰州理工大学电气工程与信息工程学院;西北民族大学数学与计算机科学学院;
【基金】：国家自然科学基金资助项目(61561042) 西北民族大学引进人才基金资助项目~~
【分类号】：TP391.41
【相似文献】
相关期刊论文 前6条
1 陈华;胡春海;;基于积分图像的人体肢体检测算法[J];应用光学;2011年05期
2 韩新宇;朱齐丹;章慧君;;基于人体头肩部形状的身份识别[J];应用科技;2006年12期
3 宋涛;李林涛;宋振宇;;一种抗遮挡的全自动Mean-shift跟踪算法[J];信息安全与通信保密;2010年09期
4 黄文丽;范勇;李绘卓;薛琴;唐遵烈;李立;;改进的混合高斯算法[J];计算机工程与设计;2011年02期
5 李世锋;文志强;吴岳忠;;基于改进的GMM参数估计的目标检测方法[J];重庆工商大学学报(自然科学版);2013年05期
6 ;[J];;年期
相关硕士学位论文 前3条
1 李金奎;视频中运动目标的检测和提取方法研究[D];北京邮电大学;2011年
2 杨帆;智能化考场中考生考勤的关键技术研究[D];昆明理工大学;2014年
3 赵海鹏;视频中运动目标检测与跟踪算法的研究和实现[D];电子科技大学;2013年
，本文编号：2342432
本文链接：https://www.wllwen.com/kejilunwen/ruanjiangongchenglunwen/2342432.html
