基于机器视觉的注射泵注射器液位检测  

Liquid level detection of syringe pump and syringe based on machine vision

作　　者：刘云晞 郭淼 高晶敏[1] 李越 LIU Yunxi;GUO Miao;GAO Jingmin;LI Yue(School of Automation,Beijing Information Science and Technology University,Beijing 100192,China)

机构地区：[1]北京信息科技大学自动化学院,北京100192

出　　处：《传感器与微系统》2023年第8期139-142,147,共5页Transducer and Microsystem Technologies

基　　金：北京信息科技大学分类发展项目(2121YJPY217,2020KYNH226);北京信息科技大学校基金资助项目(2025015);民用空间高级研究项目(D030105)。

摘　　要：提出一种检测注射泵输液的方法,以减少护士与患者的接触频率,减轻护士的劳动强度。利用机器视觉技术,结合Bilateral filter和Prewitt算子,对采集到的注射泵图片进行处理,获得准确的液位位置。图像由工业相机拍摄。经过灰度处理和图像滤波后,利用Prewitt边缘检测算法得到清晰的边缘图像,再利用Shi-Tomas角点检测算法和亚像素坐标显示检测注射器液位,无需外部辅助测量设备。通过精确定位注射泵的液位,可以远程获取注射泵输液的实时流量和药液剩余量,从而提高护士的工作效率。对50 mL注射器的26幅不同液位图像进行多次测试后,测试结果表明:该相对误差为-2.625%~4%。26张图片绝对误差均在±0.58 mL以内,有25张图片的检测结果绝对误差在±0.5 mL以内。A method to detect the infusion of a syringe pump is proposed to reduce contact frequency between nurses and patients and reduce the labour intensity of nurses.Using machine vision technology,combined with Bilateral filters and Prewitt operators,the collected pictures of the syringe pump are processed to obtain an accurate liquid level position.The image is taken by an industrial camera.After grayscale processing and image filtering,the Prewitt edge detection algorithm is used to obtain a clear edge image,and then the Shi-Tomasi corner detection algorithm and sub-pixel coordinate display are used to display the detected liquid level of the syringe without external auxiliary measuring equipment.By accurately positioning the liquid level of the syringe pump,the real-time flow rate of the syringe pump infusion and the remaining amount of liquid medicine can be obtained remotely,thereby improving the work efficiency of the nurse.Multiple tests are performed on 26 different liquid level images of a 50 mL syringe.The test results show that the relative error is-2.625%~4%.The absolute errors of 26 pictures are all within±0.58 mL.The absolute errors of the detection results of 25 pictures are within±0.5 mL.

关 键 词：机器视觉 Prewitt算法 液位检测 角点检测 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术] TP274.2[自动化与计算机技术—计算机科学与技术]