摘要
视觉检测由于其非接触和检测速度快、检测精度高等多方面的优势,被广泛用于工业检测的诸多场合。由于光晕现象的存在,应用视觉检测的方法在进行图像尺寸测量时,往往会带来较大的检测误差。设计了一种用于消除光晕现象影响的视觉算法,具有鲁棒性高,重复精度好等优点。通过对自制的等间距圆环的圆形靶标进行多种曝光环境下的目标尺寸检测,对直接应用大津法进行图像阈值分割和应用算法消除光晕现象之后再计算的靶标外圆环直径结果进行多个方向的误差分析对比。实验结果证明该算法具有较强的鲁棒性,能够适用于多种光照环境,并且检测的精度能够满足大部分场合的应用需求。
Visual inspection is widely used in many occasions of industrial inspection due to its advantages of non-contact,fast inspectionspeed and high inspection accuracy.Due to the existence of the halo phenomenon,the application of visual inspection methods in imagesegmentation will often bring greater detection errors.A visual algorithm for eliminating the halo phenomenon was designed,which had theadvantages of high robustness and good repeat accuracy.The self-made equal-spaced circular targets were used to segment images in multipleexposure environments,and the results of the target spacing calculated after the direct application of the Otsu method for image thresholdsegmentation and the application of algorithms to eliminate the halo phenomenon were compared by standard deviation analysis.The resultproves that the algorithm has strong robustness,can be applied to a variety of lighting environments,and the detection accuracy can meet theapplication requirements of most occasions.
作者
谢柳辉
冯晓蕾
周晓
王进举
李新成
Xie Liuhui;Feng Xiaolei;Zhou Xiao;Wang Jinju;Li Xincheng(Guangdong Institute of Special Equipment Inspection and Research Dongguan Branch,Dongguan,Guangdong 523000,China;School of Mechanical and Electronic Engineering,Wuhan University of Technology,Wuhan 430000,China)
出处
《机电工程技术》
2021年第6期46-47,82,共3页
Mechanical & Electrical Engineering Technology
基金
广东省质量技术监督管理局科研项目(编号:2018CT07)。
关键词
光晕现象
图像分割
鲁棒性强
高精度
halo phenomenon
image segmentation
strong robustness
high precision
