摘要
针对三帧差分法不能够实现将复杂场景中的位置不变运动目标与位置变化的运动目标分开,本文提出了一种三帧差分法的改进方案,该方案首先利用三帧差分法检测出场景中所有运动的物体,然后利用背景辅助识别的方法将位置不变自身变化的运动物体识别出来。实验证明该方法可以有效地减少三帧差分法存在的误判现象。
The normally used method, three-frame difference, can not function properly in detecting moving objects from a complicated fixed background. This paper presents an improved scheme on the three frame difference. The scheme firstly detects out all moving objects from complicated background by making use of three frame difference method, and then distinguishes moving objects based on analysis of background information. Experiment results show that the presented method may effectively decrease detecting mistakes produced by three frame difference method.
出处
《山东建筑大学学报》
2007年第6期557-560,共4页
Journal of Shandong Jianzhu University
关键词
目标检测
图像分割
三帧差分
视频监控
object detection
image segmentation
three frame difference
video surveillance
