摘要
某些远程网络监控系统,以指针仪表或数字仪表为主要监控对象,监控图像的主要特点是仅仪表指针或数码显示存在显著变化,背景变化很少。利用该特点,采用卡尔曼滤波更新背景、高阶统计分析处理含噪差值图像两者相结合的方法来提取准确的指针位置或数码显示,并给出目标区域的坐标。这样可在编码端只压缩并传输提取目标,到远程解码端再与背景合成得到实时图像,以减少传输量提高监控实时性,同时为实现监控的智能化做准备。仿真结果表明,该方法可获得较准确的指针位置及数码显示。
In some remote monitoring system, pointer or digital instruments are the main objects. In this kind of monitoring system, the pointer or number changes a lot, while the background changes a little. The method of combining Kalman Filter used to update background with high order statistics used to process noise difference is adopted to get the position of the pointer or the value of the number. And then the coordinate of the target area is presented. So moving ob- ject images on site are compressed and transfered, then inserted to the background for later intelligent recognizing and processing. This mode can not only improve the real time capability by reducing the amount of the transmission data but also lay a solid foundation for monitoring intellectualization. The simulation result indicates that this method can extract the pointer position and digital value precisely and effectively.
出处
《仪表技术》
2010年第10期42-45,共4页
Instrumentation Technology
关键词
背景更新
卡尔曼滤波
前景获取
高阶统计分析
background updating
Kalman filter
foreground extraction
high order statistics
