摘要
为解决人工检测电镀卫浴产品表面划痕存在的漏检、效率低等问题,提出基于机器视觉的电镀卫浴产品表面划痕检测系统。利用工业相机采集电镀卫浴产品缺陷图像,通过调整相机和产品的夹角并设计暗室照明环境解决电镀卫浴产品成像倒影、局部曝光问题;通过高斯滤波算法去除背景噪声,采用正弦带通滤波器和傅里叶变换将图像从空间域转换频域后经过阈值分割提取缺陷区域。实验结果表明,划痕检测系统能够检测出表面划痕长度在0.4 mm以上的缺陷,准确率达到96%,检测时间控制在350 ms左右,满足检测要求。
In order to solve omission,low efficiency and other problems in manual inspection of surface scratches on electroplating sanitary product,a surface inspection system for electroplating sanitary products based on machine vision was proposed.Images of defective electroplating sanitary products were captured by industrial cameras,and the angle between the camera and the product was adjusted and a darkroom lighting environment created to conduct imaging reflection and partial exposure.Gaussian filter algorithm was used to remove background noise,the sinusoidal bandpass filter and Fourier transform used to transform the image from the spatial domain to the frequency domain,and threshold segmentation performed to extract the defect region.The experimental result shows that the scratch detection system can detect a surface scratch 0.4 mm long and longer with a 96%accuracy rate within 350 ms,which satisfies the industrial testing requirements.
作者
张洪铭
葛晓宏
李辉
ZHANG Hongming;GE Xiaohong;LI Hui(School of Mechanical&Automotive Engineering,Xiamen University of Technology Xiamen 361024,China)
出处
《厦门理工学院学报》
2019年第5期8-13,共6页
Journal of Xiamen University of Technology
关键词
电镀卫浴产品
表面划痕检测
机器视觉
傅里叶变换方法
正弦带通滤波器
频域
electroplating sanitary products
surface scratch detection
machine vision
Fourier transform
sinusoidal bandpass filter
frequency domain
