医用输液生产中药品安全隐患分析及对策  

Analysis of Potential Drug Safety Problems in Production of Medical Infusion and Their Countermeasures

作　　者：杨福刚[1] 

机构地区：[1]山东工商学院信息与电子工程学院

出　　处：《中国安全科学学报》2010年第5期102-108,共7页China Safety Science Journal

基　　金：国家自然科学基金资助(60673153;60970105);山东省自然科学基金资助(Y2007G22)

摘　　要：针对医用输液生产中易受到异物污染、存在安全隐患的现状,分析安全隐患产生的主要环节及成因。当前生产中普遍采用的人工灯检方法,由于劳动强度大、工人易疲劳,并不能从源头上完全杜绝含有异物微粒的输液进入临床。为此,提出利用机器视觉技术对输液中异物微粒进行在线自动检测的解决方案。首先建立输液中异物粒子的运动轨迹数学模型;然后提取输液视觉图像序列中每个可能目标的有效特征,通过特征匹配和轨迹关联进行甄别检测;最后对视觉系统进行标定,确定检测到的异物微粒的粒径大小及微粒数量。实验表明,该技术对输液中微小异物检出确率达95%以上,能从源头上有效避免医用输液生产中异物污染。In view of the status quo of medical infusion being susceptible to contaminations and existing potential safety problems in its production, main causes and producing process for potential safety problems are analyzed. The commonly-used manual inspection, due to causing high labor intensity and the fatigue of workers, can not completely prevent unqualified infusion into clinical field from the origin. So, a novel particle detection technology based on machine vision which can automatically make an on-line detection of foreign particles in infusion producing is proposed. Firstly, the mathematical model of particles＇ centrifugal motion trajectory is established. Then, by extracting the effective features of potential targets from image sequences, matching these features of adjacent frames and associating the potential motion trajectories of targets, the existence of the particles can be determined. The final calibration of the vision system identifies the size and the number of the particles detected. Experiments show the detection rate of foreign particles in infusion producing can reach more than 95% with the proposed technology, which can effectively prevent the contamination of medical infusion.

关 键 词：医用输液 安全隐患 轨迹关联 机器视觉 微粒检测 

分 类 号：X913.4[环境科学与工程—安全科学]