[1]沈云青,熊卫华,黄为民,等.基于模板匹配与 Hough 圆检测的仪表识别[J].计算机技术与发展,2021,31(04):69-73.[doi:10. 3969 / j. issn. 1673-629X. 2021. 04. 012]
　SHEN Yun-qing,XIONG Wei-hua,HUANG Wei-min,et al.Instrument Recognition Based on Template Matching andHough Circle Detection[J].,2021,31(04):69-73.[doi:10. 3969 / j. issn. 1673-629X. 2021. 04. 012]
点击复制

基于模板匹配与 Hough 圆检测的仪表识别()
分享到：

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:
31
期数:
2021年04期
页码:
69-73
栏目:
图形与图像
出版日期:
2021-04-10

文章信息/Info

Title:
Instrument Recognition Based on Template Matching andHough Circle Detection
文章编号:
1673-629X(2021)04-0069-05
作者:
沈云青; 熊卫华; 黄为民; 许 伟
浙江理工大学 机械与自动控制学院,浙江 杭州 310018
Author(s):
SHEN Yun-qing; XIONG Wei-hua; HUANG Wei-min; XU Wei
Faculty of Mechanical Engineering & Automation,Zhejiang Sci-Tech University,Hangzhou 310018,China
关键词:
仪表定位; 指针识别; K-means 二值化; 模板匹配; Hough 圆检测
Keywords:
instrument positioning; pointer recognition; K-means binarization; template matching; Hough circle detection
分类号:
TP391
DOI:
10. 3969 / j. issn. 1673-629X. 2021. 04. 012
摘要:
随着自动化技术的不断发展,国内电力系统的自动化程度不断提高,无人变电站、无人配电房也逐渐普及。 针对变电站中指针式圆形仪表的识别, 提出了一种基于模板匹配与 Hough 圆检测的仪表识别方法。首先对图像进行预处理,将摄像头获取到的 RGB 彩色图像进行灰度化以及图像的局部直方图均衡化,将预先采集的仪表模板与待测图片进行两次模板匹配,获取仪表大致位置并对图像进行定位裁剪,通过 Hough 圆检测获取圆形表盘位置及其圆心坐标,将图像通过K-means方法二值化,再通过旋转虚拟直线法对指针进行拟合识别,通过角度与刻度的对应关系计算读数。 利用 Python 和OpenCV 视觉库实现算法,实验结果表明,该算法对于指针式圆形仪表识别效果显著,定位表盘位置准确,指针识别和角度计算有较高的精度。
Abstract:
With the continuous development of automation technology,the degree of automation of China爷 s power system continues to increase,and unmanned substations? ? ? and unmanned distribution rooms are also gradually popularized. Aiming at the identification of pointer-type circular meters in substations, an instrument identification method based on template matching and Hough circle detection is proposed. First, the image is preprocessed, and the RGB color image obtained by the camera is grayed and the local histogram equalization of the image is carried out. The pre-collected instrument template is matched with the picture to be tested twice to obtain? ? the approximate position of the meter,and then the image is positioned and clipped. The position of the circular dial and its center coordinates are obtained by Hough circle detection. The image is binarized by K - means method,and then the pointer is fitted and identified by rotating virtual line method. The reading is calculated by the corresponding relationship between angle and scale. The algorithm is realized by Python and OpenCV vision library. The experiment shows that the proposed algorithm has a significant effect on the pointer-type circular instrument recognition,the positioning dial position is accurate,and the pointer recognition and angle calculation have high accuracy.
更新日期/Last Update: 2020-04-10