摘要
针对小模数齿轮尺寸限制带来的检测困难问题,采用目前精密测量中流行的机器视觉检测方法 ,搭建了一套快速检测小齿轮缺陷以及几何参数的系统。研究各向异性扩散法在齿轮图像预处理过程中的应用;结合Otsu阈值和图像形态学对齿轮边缘轮廓区域进行快速定位;采用Deriche边缘提取算法实现对小齿轮边缘的快速提取;基于形态学处理实现对齿轮的缺陷检测;采用辅助圆扫描法测量齿轮的齿距。实践证明,该检测系统能达到对小齿轮缺陷以及几何参数的高效、实时、精确检测的要求。
For detection difficulties small modulus gear size limits brought, using machine vision inspection method which is popular in precision measurement to build a rapid gears detection system. Application of anisotropic diffusion method in image pre-processing is studied; Combines Otsu threshold and image morphology to quickly locate gear edge contour area. Uses edge extraction operator called ＇Deriche＇ to quickly achieve gear edge contour. Achieves defection of gear based on morphological processing. Measures gear tooth pitch using auxiliary circle scanning. Practice proves that the detection system can achieve the goal to detect small gear efficiently and accurately.
作者
周鑫
ZHOU Xin(School of Mechatronic Engineering, East China Jiaotong University, Nanchang 330000 Chin)
出处
《自动化技术与应用》
2017年第10期104-107,127,共5页
Techniques of Automation and Applications
