基于背景建模的动态场景目标检测

背景建模一直是运动目标检测中的一个重要课题。该文提出一个适用于动态背景的基于非参数估计的前景背景对比模型。模型通过核函数估计的方法模拟了像素点五维特征向量(彩色灰度值，图像坐标)的概率分布，并在图像序列中滚动更新。对于每一个新入帧通过马尔可夫随机场最大后验概率判决框架将前景背景全局分割问题转化为最大流最小切求解。实验证明，上述算法能够在一般目标检测，特别是动态场景(摇动树枝等)的检测中取得较好的效果。

Object Detection in Nonstationary Scenes Based on Background Modeling

Background modeling is an important issue in accurate detection of moving objects. This paper presents a novel non-parametric foreground-background model which explores the complex temporal and spatial dependencies in nonstationary scenes. The model estimates the probability of observing pixels’ five-dimensioned feature vector which represents its intensity values and spatial position information. The model is built and rolling-updated by kernel density estimation. And a Maximum A Posteriori-Markov Random Field(MAP-MRF) decision framework is used to segment the foreground and background by solving a graph-cut. Extensive experiments with nonstationary scenes demonstrate the utility and performance of the proposed approach.