以仪表结构为先验特征的指针式仪表图像识别

在指针式仪表图像识别方法中,为克服的指针定位易受干扰的问题,文中基于Canny边缘算子和Hough直线检测,提出了一种以表盘半径和指针的旋转中心作为先验特征的指针定位算法。首先以旋转中心为中点,表盘直径为边长,提取正方形的感兴趣区域以减小Canny算子的作用时间;其次将Hough变换后的所有直线与结构特征进行比较,剔除不满足距离和位置条件的干扰直线;最终结合Canny边缘算子特点,分析并修正数值计算的误差。

Pointer Instrument Image Recognition Based on Priori Characteristics of Instrument Structure

In view of the fact that the pointer positioning is prone to interference by pointer instrument image recognition method,an improved pointer algorithm is proposed based on Canny edge operator and Hough line detection with the center of rotation and the radius of the dial as priori characteristics.With the center of rotation as the midpoint and the dial diameter as the side length,the square region of interest is extracted in order to reduce the duration of action of the Canny operator,followed by a comparison between all the straight lines after the Hough transform and structural characteristics to exclude the lines that do not meet the distance and position requirements.Finally,based on the Canny edge operator characteristics,the error of numerical calculation are analyzed and corrected.