基于卡尔曼滤波的液位检测与跟踪  

Moving Liguid-line Detecting and Tracking Based on Kalman Filter

作　　者：陈金丽[1] 张帆[1] 张显[1] 

机构地区：[1]湖南大学电气与信息工程学院,湖南长沙410012

出　　处：《计算机仿真》2009年第9期317-320,共4页Computer Simulation

基　　金：湖南省自然科学基金(2007JT1024)

摘　　要：液位是工业测量的重要参数之一,如何准确有效的对液位进行测量一直是研究的一个重点。为解决上述问题提出将视频技术应用于液位测量的新方法。当出现干扰时,以往的直接测量算法很难准确测量,而运用此方法,首先根据液位运动特性,建立其运动的数学模型,在采用更新函数的背景差法基础上,提出了一种卡尔曼滤波的预估校正算法,避免了测量中对单帧图像信息的依赖,提高了检测系统抗扰性。实验仿真表明,算法有效地保证了液位跟踪的准确性、稳定性和连续性。Liquid level is an important parameter of commercial measure. How to measure the liquid level accurately and effectively has been a focus of study. A new method of using video technology in liquid level measurement was proposed. The previous algorithm failed to measrue it accurately when there was interference, but using this method, first of all, based on the liguid moving characteristic, a moving mathematical model was established. Taking the background difference method for fusing updating functions as the basis, an advance revising algorithm based on Kalman filter was proposed, which avoided the depending on single image information, and improved the anti - nuisance of detecting system. Simulation experiments showed that the method ensured the veracity and stability and continuity of liguid - line tracking.

关 键 词：液位检测 目标跟踪 扩展卡尔曼滤波 运动预测 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]