双重自适应码本模型在运动目标检测中的应用

Moving Objects Detection with Double Adaptive Codebook Model

针对经典码本模型在复杂环境下的前景检测存在自适应动态背景能力不足的问题,提出一种双重自适应码本模型算法.该算法分别对前景和背景进行建模和更新,使得前背景可以在预设参数的控制下进行相互转换,消除了背景变化造成的虚影现象;根据命中次数对模型码本中码字的位置进行快速冒泡排序,提高了活动码字首次匹配成功的概率;利用短时滑动窗口的方法对像素变化信息进行存储,实现了像素的均值及偏差的快速实时跟踪,解决了动态背景的模型自适应问题.实验结果表明,文中算法具有较好的检测效果和实时性能,适用于复杂环境条件下的前景目标检测.

A dual adaptive codebook model algorithm is proposed is this paper to improve the adaptive capacity of moving object detection in dynamic background. The algorithm models and updates the foreground and background respectively. A parameter controlled converter is employed to achieve mutual conversion between the foreground and background model, which can effectively remove the ghost produced by the partial change of background. A rapid bubble sorting algorithm is proposed to sort the position index of the code words according to their hit times, which can improve the probability of matching the active code word at the first time. A short sliding time window is used to buffer the change of the pixel and trace the average and deviation, which can effectively solve the adaptive problem of model update with dynamic background. Experiment results show that the improved algorithm has adequate detection accuracy and real-time performance and it is suitable for moving object detection in complex environment conditions.