基于质心迭代算法的安瓿药液杂质跟踪检测

针对基于旋转急停法的安瓿药液杂质跟踪检测效率和准确性低的问题,提出一种采用质心迭代算法实现运动目标的跟踪设计方法。利用交互方差系数估计背景提取运动目标,采用非核非参数的质心迭代算法对运动目标进行跟踪,并根据运动目标的轨迹排除气泡干扰,判断杂质的存在情况。对算法基于Matlab平台进行仿真,结果表明:此算法具有高效、高准确率、易实现等特点,适用于安瓿药液杂质的在线检测。

Impurities Tracking and Detecting of Ampoule Liquid Based on Centroid Iteration Algorithm

This paper proposes a method for tracking the moving targets based on centroid iteration algorithm to improve the efficiency and accuracy of tracking and detecting the impurities in ampoule liquid. By estimating the background with the interaction coefficient of variance to extra the moving objects, they are tracked with the centroid iteration algorithm, which are non-nuclear and non-parametric. The existence of impurities is judged based on the trajectory of the moving objects by excluding the interference of the bubbles. The simulation carried out on Matlab shows that the algorithm is high efficiency, high accuracy and easy to implementation, which is suitable for ampoule liquid impurity on-line detection system.