Abstract: In order to reduce the complexity of the background extraction algorithm in time and space, this paper presents a background extraction method which is combination of differential image block, background subtraction and frame difference method. Differential image block is classified. A sub-block classification method of statistics of pixel values is presented. Background to different types of  block with different update policy is updated. The algorithm is effective to solve the problem about the influence that the update process moving objects stay, background objects stay out. The experimental results show that the algorithm has the characteristics of fast operation and high robustness and it can extract real-time background accurately.

Key words: image block, frame-difference, background subtraction, background extraction

摘要： 为了降低背景提取算法的时间复杂度和空间复杂度，提出一种结合差分图像分块、背景减除和帧间差法的背景提取方法。对差分图像进行分块分类，提出了一种统计像素值的子块分类法，对不同类的块用不同的更新策略进行背景实时更新。该算法有效解决了背景更新过程中运动目标逗留、背景物体移入移出等问题的影响。实验结果表明该算法运算速度快、鲁棒性高、能准确地提取实时背景。

关键词: 图像分块, 帧间差分, 背景减除, 背景提取