摘要
设计了一种基于机器视觉的温控器产品外观检测装置,检测温控器编码有无、绝缘纸长度、绝缘纸不对称、长短爪等参数。应用迭代法实现了二值化阈值的确定,应用Canny算法实现产品的边缘检测,提取产品图像的像素坐标,实现外观检测。实验表明,系统的不良品检出率为100%,良品误检率低于0.26%。
This paper designed a device to detect the appearance parameters of the temperature controller based on machine vision, which realized the detection of the coding of temperature controller, the length and asymmetry of insulating paper and the parameters of the claw. We used iterative method to ascertain the binarization threshold and Canny algorithm to realize the image＇ s edge detection, then, the pixel coordinate of the image could be extracted to realize the appearance detection. Experiments show that the detection rate of defective products is 100%, while the false drop rate of accepted products is lower than 0.26%.
出处
《智慧工厂》
2017年第12期80-82,共3页
Smart Factory
基金
2015年江苏省"六大人才高峰"资助项目
项目编号:2015-ZBZZ-038
2016年省科技厅产学研前瞻性联合研究项目
项目编号:BY2016065-12
关键词
温控器
外观检测
机器视觉
图像处理
Temperature controller Appearance detection Machine vision Image processin
