基于连续帧间差分和背景差分法在输液速度检测中的研究  

Continuous frame difference and background subtraction method based on the detection of infusion speed

作　　者：袁伟[1,2] 王金明[1] 吴加明[1] 秦水介[2] 

机构地区：[1]贵州大学大数据与信息工程学院,贵州贵阳550025 [2]贵州省光电子技术及应用重点实验室,贵州贵阳550025

出　　处：《电子设计工程》2015年第14期176-179,共4页Electronic Design Engineering

摘　　要：为了方便快速实时的检测出医疗输液过程中的液滴速度,本文将基于连续帧间差分与背景差分的算法相融合运用到液滴速度检测中,该算法利用Matlab工具对输液视频提取出每帧的图像序列。对图像序列首先进行连续帧间差分,然后对差分图像进行滤波处理获取背景图像,之后连续帧差图像与背景图像进行差分,再利用阈值处理即可准确的检测出液滴的整个过程,通过分析液滴的整个过程,本文选取液滴滴落液面的过程作为液滴速度检测的依据,统计其运动像素点,取合适的阈值计算出液滴速度为34滴/分,与实际速度相符合。In order to facilitate rapid real-time detection of medical infusion infusion speed,This article uses intergration of continuous frame difference and background subtraction algorithm applied to the detection of infusion speed,The algorithm uses Maflab tools infusion video extracted from the image of each frame of the sequence.First,treatment of image sequence for continuous frame difference.Acquiring background image by difference image filtering,After continuous frame difference image subtracting background image.Finally, using a threshold value processing to detect accurately the entire process of the droplet. In this paper, the process of selecting dripping liquid droplet velocity as the basis for detection,statistical its motion pixels, calculate the velocity of the droplet 34 drops/min,consistent with the actual speed.

关 键 词：连续帧间差分 背景差分 运动像素点 液滴速度 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]