摘要
当前呼出气体酒精含量检测仪检定工作耗时耗力,检定过程需要人工进行判读和记录,针对该问题研制一套呼出气体酒精含量检测仪自动检定装置。依据检定规程的要求,设计配气装置实现对气体的气路设计及温度和流量等的控制。基于机器视觉系统使用Halcon联合C#开发识别算法,利用Canny边缘算法提取感兴趣区域,定位提取示值字符后训练多层感知器来识别示值结果,并实现检定过程的自动选择和报告的自动生成。实验结果表明:对五种型号的呼出气体酒精含量检测仪示值的识别准确率达到了99.5%以上,平均识别时间小于200 ms,对后续检定的平均全过程时间从3.5天缩短至1.5天。该装置满足检定规程的要求,可以应用于各地计量机构的检定工作中,具有实用价值。
The current verification work of exhaled gas alcohol content detectors is time⁃consuming and labor⁃intensive,and the verification process requires manual interpretation and recording.To address this issue,an automatic verification device for exhaled gas alcohol content detectors has been developed.According to the requirements of the verification regulations,a gas distribution device is designed to achieve gas circuit design and control of temperature and flow rate.On the basis of the machine vision system,Halcon and C#are combined to develop the recognition algorithm,Canny edge algorithm is used to extract the region of interest(ROI),and multilayer perceptron(MLP)is trained to recognize the indication results after locating and extracting the indication characters,and the automatic selection of verification process and automatic generation of reports are realized.The experimental results shows that the recognition accuracy of the five types of exhaled gas alcohol content detectors reaches over 99.5%,the average recognition time is less than 200 ms,and the average entire process time for subsequent verification is shortened from 3.5 days to 1.5 days.The designed device can meet the requirements of the verification regulations and can be applied to the verification work of various metrological institutions,so it has practical value.
作者
张湖波
邹细勇
邵建文
赵存彬
骆蕾
ZHANG Hubo;ZOU Xiyong;SHAO Jianwen;ZHAO Cunbin;LUO Lei(College of Optical and Electronic Technology,China Jiliang University,Hangzhou 310018,China;Zhejiang Institute of Metrology,Hangzhou 310018,China)
出处
《现代电子技术》
北大核心
2024年第1期111-117,共7页
Modern Electronics Technique
基金
浙江省市场监督管理局科研项目(20210173)。
