摘要
运动目标检测是视频处理的基础,而目标的阴影在很大程度上影响了目标的真实形状,干扰了真实目标的检测。本文提出了一个以混合高斯模型为基础,结合多特征的运动目标检测方法。将阴影消除算子、帧差、方差及彩色信息融合到背景模型中,能较准确地检测运动目标并消除阴影的影响。
Motion detection is the basis of video processing. The shape ot an object is attected mostly by xts shadow, by which the true object detection is interfered. In this paper, a multi-feature motion detection method is put forth based on the hybrid Gaussian model. By putting the shadow elimination operator, frame difference, variance and color information into the background model, a moving object can be detected more correctly by the method and the effect of shadow can be eliminated.
出处
《计算机工程与科学》
CSCD
2007年第8期40-42,共3页
Computer Engineering & Science
关键词
混合高斯模型
运动检测
阴影
hybrid Gaussian model
motion detection
shadow
