指针式仪表读数的机器视觉智能识别方法  

Intelligent recognition method for reading of pointer instrument based on machine vision

作　　者：施滢[1] 夏春华[2] 胡琳娜[1] 孟迎军[1] 

机构地区：[1]南京理工大学紫金学院,江苏南京210023 [2]农业部南京农业机械化研究所,江苏南京210014

出　　处：《传感器与微系统》2017年第11期47-49,52,共4页Transducer and Microsystem Technologies

基　　金：中央级公益性科研院所基本科研业务费专项项目(S201715)

摘　　要：为快速准确地自动识别指针式仪表读数,采用机器视觉技术,结合减影法和Hough变换法对仪表读数进行智能识别。对指针式仪表图像进行图像二值化、形态学处理和边缘检测预处理;利用Hough变换检测仪表中的指针,计算得到指针方向和定位圆形,实现指针式仪表的智能识别。实验结果表明:读数识别的平均相对误差为0.91%,精度较高,能较好地识别指针式仪表读数。In order to identify the reading of pointer instrument automatically and accurately,an intelligent recognition method of pointer instrument is proposed based on machine vision. The reading of pointer instrument is recognized based on machine vision,combined with image subtraction and Hough transformation. Pointer instrument image is preprocessed by image binarization,morphological processing and edge detection. Direction of the pointer and the position of the circle are determined by Hough transformation,in order to realize intelligent recognition method of pointer instrument. Experimental result shows that the average relative error is 0.91 %,the precision is high,the reading of pointer instrument can be recognized accurately by this method.

关 键 词：指针式仪表 机器视觉 智能识别 HOUGH变换 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]