一种自适应码书模型背景更新算法

Adaptive Codebook Model Background Update Algorithm

因背景更新过程中运动信息不足,造成在处理缓慢移动目标和只有局部运动目标时常常发生误判,为解决上述问题,通过提取运动目标的空间整体信息,提出了一种自适应的码书模型背景更新算法。该方法通过对运动目标空间信息变化进行分析,寻找前景中潜在背景,然后联合像素时域统计信息,得到真正的背景模型。实验结果表明,该算法可以快速适应背景变化,能明显减少对运动信息不足目标的误判,同时保证目标检测的完整性。

Codebook model is a widely used method for video object detection and its performance is dependent on the quality of background model. Existing update algorithm can＇t work well when the object motion is sparse and insufficient. They lead to misjudgments in processing slow-moving objects and only the local moving object. To solve the problem,a new method was presented. The method gets the moving object spatial information and then joint pixel time-domain statistical information to complete background update. Experimental results show that this algorithm not only adapts to environ- mental changes but also can handle the limited object motion and ensure the integrity of the moving object.