基于ECO跟踪的输液瓶可见异物检测识别方法

Detection and recognition of impurity in infusion bottle based on ECO tracking

输液瓶作为常见的注射药剂,研究其药液内是否存在异物是安全和质量检测领域的重要课题。可见异物自动检测设备通常采用旋转一急停的方式使瓶中异物运动旋转,采集多帧图像序列,利用机器视觉的方法来判断液体中是否存在异物,而这种方式会产生大量气泡,严重干扰检测精度。提出一种基于ECO跟踪的输液瓶可见异物检测方法,以图像序列的感兴趣区域为对象,借助帧差法找到图像序列中最可能存在异物的图像作为起始帧图像,并找到起始帧图像中的运动目标,建立初始模板,利用E C O算法跟踪运动目标,记录其运动轨迹,分析运动特性以此辨识目标是否为异物。实验结果表明,提出的方法可以有效地检测识别异物,去除气泡干扰,并且具有较高的识别率和较低的漏检率,可以达到工业检测的要求。

Infusion bottle is a kind of common injection drug.It is an important subject in the field of safety and quality detection to study whether there are foreign insoluble particles in the infusion bottle.The impurity automatic detection equipment usually works by rotating-stopping mode,collecting multi-frame image sequence,and using machine vision method to judge whether there is impurity in the liquid.This approach will produce a large number of bubbles,seriously interfering with the detection accuracy.This paper presents a method for detecting impurity in infusion bottles based on ECO tracking.The region of interest in the image sequence is taken as the object image,and the image with foreign particles in the sequence is found by frame difference method as the start-tracking image.Then we can obtain the moving object in the start-tracking image to establish the initial template.The ECO algorithm is used to track the moving object,record its trajectory and analyze its motion characteristics,to identify whether the target is a foreign particle.The experimental results show that the proposed method can effectively detect and identify impurity,remove bubble interference,and has a high recognition rate and a low leakage rate,which can meet the requirements of industrial detection.