摘要
机械式仪表因精度高、读数方便、可调等优点在各行业中被广泛使用。采用视觉识别的方法来对目前广泛使用的机械式仪表进行识别,可以显著提高自动化仪表示数读取的水平,同时也能提升检测准确率。系统基于Python分析常用的边缘检测、霍夫变换等图像预处理方法来识别机械仪表指针读数,并采用归一化、灰度等预处理方法,对图像进行了倾斜校正,实现读数的准确识别。通过不同表盘实验,验证了该系统能较好地识别机械表读数。
Mechanical instruments are widely used in various industries due to their high accuracy,convenient reading,and adjustability.The use of visual recognition methods to identify the widely used mechanical instruments can significantly improve the level of automatic instrument reading and also improve detection accuracy.This system is based on image preprocessing methods such as edge detection and Hough transform commonly used in Python analysis to identify mechanical instrument pointer readings.Normalization,grayscale,and other preprocessing methods are used,and the image is corrected for tilt to achieve accurate recognition of readings.Through four different dial experiments,it is verified that the system can effectively recognize mechanical meter readings.
作者
刘岩
努尔兰·吐尔达洪
张连涛
LIU Yan;NUERLAN Tuerdahong;ZHANG Liantao(Xinjiang Institute of Engineering,Urumqi 830023,China)
出处
《自动化技术与应用》
2025年第8期94-97,155,共5页
Techniques of Automation and Applications
基金
新疆维吾尔自治区教育厅科研项目(XJEDU2020Y043)。
关键词
视觉识别
自动化仪表
机械表识别
边缘检测
图像预处理
visual recognition
automatic instruments
mechanical meter recognition
edge detection
image preprocessing
