摘要
为实时智能监控变电站安全生产区域内的移动目标,克服现有视频系统人工切换图像和肉眼判断所造成的漏检和滞后问题,对变电站内运动目标的自动检测与识别跟踪技术进行了研究;基于背景差分法实现了人物动态目标检测,提出了基于颜色直方图的粒子滤波人物动态目标跟踪方法;通过提取目标颜色特征,建立目标状态模型和系统模型,进而准确定位目标;研发了变电站安全事件视频自动识别跟踪系统.系统应用结果表明:算法检测与跟踪的时间性能良好,能够快速识别目标,并准确跟踪目标运动轨迹,有效提升了全天候智能监控站内的安全生产能力.
In order to solve the undetected and lagged problem of video system caused by manual switching and visual judgment, the real-time detecting and automatic tracking technologies of substa- tion dynamic target were studied. The background difference method was used to detect the moving target, and then an automatic target tracking method using color histogram of particle filter was presented. Using color as target character, the target status model and system model were established. The next, target position was determined. Sequentially, a video automatic identification of safety event tracking system of substation was developed. The results of application show that the time performance of algorithm is excellent. And the dynamic target can be detected and tracked quickly and precisely. Furthermore, the intelligent surveillance ability of substation can be improved effectively.
出处
《北京交通大学学报》
CAS
CSCD
北大核心
2013年第4期153-157,共5页
JOURNAL OF BEIJING JIAOTONG UNIVERSITY
基金
中央高校基本科研业务费专项资金资助(2013JBM080)
关键词
图像处理
颜色直方图
目标跟踪
粒子滤波
image processing
color histogram
target detection
particle filter
