摘要
设计了一个面向海事场景的网络视频监控系统,系统从网络上获取航道和港区视频序列,进行分析和处理,对用户定义的受控区域进行实时监控,当受控区域出现运动目标时进行分级告警。针对海事场景背景复杂、干扰大等困难,提出了改进的混合高斯背景模型及运动检测方法。在目标跟踪方面,为运动区域建立颜色、形状和运动的混合模型,并通过距离矩阵和匹配矩阵判定目标和轨迹间的对应关系。经测试,该设计实现的各项功能都能满足海事部门对海事场景的实时监控。
An intelligent marine video surveillance system is introduced. The system obtains video streams of channels and ports through internet, then analyzes and processes them. The system monitors the user-defined regions real-timely and alarms when prohibited moving objects appear. Because the marine scene is complicated and interferential, a modified mixture Gaussians approach and a moving detection method are suggested. In order to track multiple objects, a mixed model including color, shape and motion for each detected region is set up. A distance matrix and a matching matrix give the correlations between newly detected objects and trajectories.
出处
《中国航海》
CSCD
北大核心
2008年第1期24-27,共4页
Navigation of China
基金
福建省科技计划重点项目(2005H037)
关键词
水路运输
智能视频监控
运动检测
目标跟踪
waterway transportation
intelligent video surveillance
moving detection
objects tracking