自适应核密度估计运动检测方法 提出一种自适应的核密度(kernel density estimation, KDE)估计运动检测算法. 算法首先提出一种自适应前景、背景阈值的双阈值选择方法, 用于像素素分类. 该方法用双阈值能克服用单阈值分类存在的不足, 阈值的选择能自适应进行, 且能适应不同的场景. 在此基础上, 本文提出了基于概率的背景更新模型, 按照像素的概率来更新背景, 并利用帧间差分背景模

-Adaptive kernel density estimation motion detection method, an adaptive kernel density (kernel density estimation, KDE) estimated that the motion detection algorithm, the algorithm first raised the prospect of an adaptive dual-threshold of the background threshold selection method for the classification of the pixels prime The method can overcome the dual-threshold with the shortcomings of a single threshold classification, the choice of threshold can be adaptive, and can adapt to different scenarios. On this basis, this paper presents a probability-based background update model, according to the probability of the pixel to update the background, and frame difference and background mode