帧间差分与背景差分相融合的运动目标检测算法

Moving Objects Detection Algorithm Based on Two Consecutive Frames Subtraction and Background Subtraction

摘要:

针对视频序列中运动目标检测进行了研究，提出了一种将帧间差分和背景差分相互融合的运动目标检测算法，首先选取一帧作为背景帧，确立每一个象素点的高斯模型；然后对相邻两帧进行差分处理，区分出变化的区域和没有发生变化的区域，没有发生变化的区域更新到背景帧中，发生变化的区域与背景模型进行拟合，区分出显露区和运动目标，将显露区以很大的更新率收入到背景帧中。该方法允许在有运动物存在的情况下进行建模，实验表明该方法准确率高，运算速度快，能满足实时检测的需要。

Abstract:

Aimed at the complexity of the current algorithm, an algorithm based on two consecutive frames subtraction and background subtraction is presented.At first, select a frame as a background. Then subtract two consecutive frames to find out moving area and background area. Update background with the background area which is detected. At last ,compare moving area with background to locate moving objection and uncovering area. Update background with uncovering area. The background model in this algorithm is obtained even if there are some moving objections. The results show that this algorithm combines the advantages of veracity and of runtime, and fit for real time detection.