摘要
本文在快速准确的原则上,仅利用视频序列中的时域信息,提出了一种运动对象的提取方法。首先利用二次帧差求交集的方法求出图像帧的帧差掩模,最后利用二次扫描的方法得到运动对象掩模,填充就可以提取出运动对象。实验证明,该方法提取效果较好,节省了处理时间。
Based on the rules of speediness and exactness, the method of moving object extraction is proposed only with temporal information. We obtain the frame difference mask by the intersection of three twice-frame-difference images, then get the moving object mask (MOM) by scanning every column and row. At last, by filling MOM with original values we can get the extracted moving object. The experiments show that the method can get a better result with less runtime.
出处
《计算机工程与科学》
CSCD
2006年第2期63-65,共3页
Computer Engineering & Science
基金
山东省自然科学基金资助项目(Y2001G05)
关键词
运动对象
运动区域
掩模
二次帧差
moving object
moving region
frame difference
mask
twice-frame-difference