论文部分内容阅读
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。
For medical transfusion production vulnerable to foreign body contamination, there are potential safety problems, analysis of the main aspects of safety hazards and causes. The current artificial lamp inspection method commonly used in production, due to labor-intensive, workers fatigue, and can not completely eliminate the source of foreign particles containing foreign body into the clinical. To this end, the paper presents a solution of using machine vision technology to automatically detect foreign particles in the infusion solution. Firstly, the mathematical model of motion trajectory of foreign particles in the infusion solution is established. Then, the effective features of each possible target in the visual sequence of the infusion solution are extracted and screened by feature matching and trajectory association. Finally, the visual system is calibrated to determine the detected foreign particles The size of the particle size and the number of particles. Experiments show that the technology of detection of small foreign body in the infusion rate was 95% or more, from the source to effectively prevent foreign contamination in medical infusion production.