摘要
在分析了现有算法在复杂背景下所存在的不足的基础上,提出了一种适用于视频监控系统的基于视频序列像素时空相关性检测的动目标检测方法.该方法首先用每一帧中像素的空间相关性检测出目标,再用序列图像中目标的时间相关性检验目标的真实性,从而最终确定是否有运动目标.试验表明,该方法能很好地检测出运动目标,并具有较强的抗干扰能力.
The deficiencies of the actual algorithm in the complex background are analyzed, and a method of moving object detection is given based on the detection of the space-time relativity of the pels in a video sequence which is used in the military video monitoring. By this method, the space (relativity of) the pels in a frame is used to detect the object firstly, and then the authenticity of the object is proved with the time relativity of the object in a video sequence, thus it can be found out whe-(ther ) the moving objects exist or not. It′s verified by the experimentation that the method can be used to detect the moving object from a video sequence well, and has a good anti-jamming ability.
出处
《海军工程大学学报》
CAS
2004年第4期93-96,101,共5页
Journal of Naval University of Engineering
关键词
数字图像处理
运动目标检测
视频监控
digital video processing
moving object detection
video monitoring
