摘要
根据武警勤务视频监控系统对实时性的要求,在混合高斯建模的基础上,提出一种基于帧间差分法的混合高斯背景建模方法。该方法通过帧间差分法把图像分为感兴趣区域和非感兴趣区域,对不同区域采用不同的更新方法,降低了资源利用率和计算复杂度,提高了背景重建的速度,从而提高了目标检测的响应速度,保证了实际应用需求。
By analyzing the principle of Gaussian mixture modeling and combined with the armed police service requirement of real- time video monitoring system, a method for mixed Gaussian background modeling,based on inter- frame difference one,is put forward in this article,which divides the image into interest areas and non interest ones by the frame difference method,employs different updating method for different areas to reduce the use of the resources and computation complexity,and increases the speed of background reconstruction to improve the response speed of the target detection and ensure the needs of practical application.
出处
《微处理机》
2016年第2期70-72,共3页
Microprocessors
关键词
混合高斯建模
减背景法
帧间差分法
自适应更新
目标检测
响应速度
Gaussian mixture modeling
Subtract background method
Frame differential method
Adaptive updating
Target detection
Response speed
