摘要
运动目标检测是智能视频监控分析的基础性环节。混合高斯建模算法(GMM)能有效应对多模态场景,自适应性高,然而显著缺点是运算量大。运用分块的思想并结合三帧差分法对混合高斯建模算法进行改进,提出运行效率更高的运行算法。实验数据证明,该算法显著提高了算法运行效率。
Moving object detection is the base of intelligent video surveillance analysis.The Gaussian Mixture Model（GMM） can effectively deal with multi-model scene,which is highly adaptive to the background. But its obvious disadvantage is that the computation is large.Under the combination of three-frame difference method,we had improved the Gaussian Mixture Model algorithm（GMM） by using the idea of block so that this paper proposes a more efficient algorithm.
出处
《工业控制计算机》
2016年第10期102-104,共3页
Industrial Control Computer
关键词
分块
混合高斯建模
三帧差分法
partitioning
Nixed Gaussian Model
three-frame difference
