基于自动量程识别的指针式仪表盘读数研究  

Research on Pointer Instrument Panel Indication Based on Automatic Range Recognition

在线阅读下载全文

作  者:宋思源 朱均超[1] 韩芳芳[1] 刘斌[1] 张明惠 SONG Siyuan;ZHU Junchao;HAN Fangfang;LIU Bin;ZHANG Minghui(School of Electrical Engineering and Automation,Tianjin University of Technology,Tianjin 300384,China)

机构地区:[1]天津理工大学电气工程与自动化学院,天津300384

出  处:《仪表技术》2025年第1期63-68,72,共7页Instrumentation Technology

摘  要:针对现有仪表读数方法只能识别预先设定量程下的仪表盘示数的问题,提出了一种能够自动识别并读取不同量程下指针式仪表盘示数的读数方法。使用YOLOv5模型检测出待测仪表盘,以去除背景干扰;采用U-Net模型分割出指针和刻度线,根据分割结果提取起始和结束刻度线附近的区域;利用PP-OCRv3模型识别量程信息;基于分割出的起始和结束量程刻度线与指针圆心的位置关系,采用角度法计算出仪表盘读数。实验结果表明,该读数方法不仅能准确识别量程,还能实现精确读数,其平均引用误差仅为0.8%。Aiming at the problem that existing instrument reading methods can only recognize instrument panel read⁃ings under pre⁃set ranges,a reading method that can automatically recognize and read pointer instrument panel readings under different ranges is proposed.The YOLOv5 model is used to detect the dashboard under test and remove background interference.The U⁃Net model is used to segment pointers and scale lines,and the areas near the starting and ending scale lines are extracted based on the segmentation results.The PP⁃OCRv3 model is used to identify range information.Based on the position relationship between the segmented starting and ending range scale lines and the center of the pointer,the instrument panel reading is calculated using the angle method.The experimental results show that this read⁃ing method can accurately identify the range and achieve precise reading,with an average reference error of only 0.8%.

关 键 词:指针式仪表盘 读数识别 量程识别 

分 类 号:TP391[自动化与计算机技术—计算机应用技术]

 

参考文献:

正在载入数据...

 

二级参考文献:

正在载入数据...

 

耦合文献:

正在载入数据...

 

引证文献:

正在载入数据...

 

二级引证文献:

正在载入数据...

 

同被引文献:

正在载入数据...

 

相关期刊文献:

正在载入数据...

相关的主题
相关的作者对象
相关的机构对象