指针式仪表读数的机器视觉智能识别方法

摘要

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.%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.