采用Jetson嵌入式平台的LCD仪表数字识别方法研究  

Research on Digital Recognition Method for LCD Meter Using Jetson Embedded Platform

作　　者：郭清达[1,2] 刘坚 李健 胡琨元 于广平[1,2] GUO Qingda;LIU Jian;LI Jian;HU Kunyuan;YU Guangping(Shenyang Institute of Automation,Chinese Academy of Sciences,Shenyang,Liaoning 110017,China;Guangzhou Institute of Industrial Intelligence,Guangzhou,Guangdong 511458,China)

机构地区：[1]中国科学院沈阳自动化研究所,辽宁沈阳110017 [2]广州工业智能研究院,广东广州511458

出　　处：《计算技术与自动化》2025年第1期102-106,117,共6页Computing Technology and Automation

基　　金：中国博士后科学基金资助项目(2023M730779)。

摘　　要：当前各行各业面临着数字化转型,亟需工厂各个环节设备的仪表数据进行全流程的感知,然而大量接口封闭和无数据接口的LCD数字仪表难以快速地获得数据。针对LCD仪表数字检测识别过程中的图像信息冗余、显示屏帧切换导致的数字重影和边缘侧快速部署需求,提出了一种采用Jetson嵌入式平台的LCD仪表数字识别方法。该方法通过设定感兴趣区域(Region of Interest,ROI)减少冗余信息,并结合帧间差分获得ROI阈值信息,过滤掉仪表重影图像,最后应用PP-OCRv2轻量化网络模型中的检测网络模型和识别网络模型。测试结果表明,在Jetson嵌入式平台部署程序,采用帧间差分法能够显著地过滤掉LCD仪表数字重影图像,获得稳定的仪表数值,其检测速度平均为0.37 s,满足工厂场景实际需求。At present various industries are needing digital transformation and there is an urgent need for full process perception of meter data from various equipment in factories.However,a large number of LCD digital meter with closed interfaces and no data interfaces are difficult to quickly obtain data.Due to the redundancy of image information,digital ghosting caused by screen frame switching and rapid deployment requirements on the edge side,we proposed a digital recognition method for LCD meter using the Jetson embedded platform.The method reduces redundant information by setting regions of interest(ROI),and filters meter ghosting images by setting ROI threshold value obtained through interframe difference.Finally,detection network model and recognition network model in the PP-OCRv2 lightweight network model are applied.The test results show that the program is easily deployed on the Jetson embedded platform,filter out LCD meter digital ghosting images and obtain right meter display values.The digital recognition method has an average detection speed of 0.37 s and meet the actual needs of industry scenarios.

关 键 词：数字仪表识别 帧间差分 深度网络模型 嵌入式平台 

分 类 号：TP242.2[自动化与计算机技术—检测技术与自动化装置]