文献类型：期刊文献

中文题名：改进高斯混合模型的视频运动目标检测

英文题名：Moving object detectingbased on improved mixture Gaussian

作者：杨佳

第一作者：杨佳

机构：[1]贵州理工学院信息网络中心

第一机构：贵州理工学院信息网络中心

年份：2015

卷号：0

期号：6

起止页码：147-150

中文期刊名：激光杂志

外文期刊名：Laser Journal

收录：北大核心:【北大核心2014】；

基金：贵州教育厅项目(20140023)

语种：中文

中文关键词：运动目标;视频序列;高斯混合模型;背景更新

外文关键词：moving object;;video series;;mixture Gussian;;background subtraction

摘要：视频运动目标检测具有十分重要的应用价值,其核心问题是目标检测算法,针对当前视频运动目标检测算法检测性能较差的问题,提出了一种改进高斯混合模型的视频运动目标检测算法。首先对传统高斯混合模型更新率进行改进,以适应外界环境的变化,然后引入三帧差分法实现了对背景模型的选择性更新,最后采用仿真实验测试算法的性能。实验结果表明,本文算法实现简单,可以对视频运动目标进行完整检测,相对于传统高斯混合模型具有更强的鲁棒性,而且检测速度更快。
It has very important application value of video moving object detection,the core issue is to target detection algorithm,in view of poor performance in the current video moving object detection algorithm,this paper proposes a video moving object detection algorithm based on improved Gauss mixture model algorithm. Firstly,updating rate of the traditional Gauss mixture model was improved to adapt to the changes in the external environment,and then three frame difference method is introduced to achieve the selectivity for background model updating,finally the simulation experiment is adopt to test the performance of algorithm. The experimental results show that,the proposed algorithm is simple,can realize the detection of video moving objects completely,and has stronger robustness than the traditional Gauss mixture model and faster detection speed.