基于机器视觉的斜齿轮参数检测

Helical Gear Parameters Detection Basedon Machine Vision

为了实现斜齿轮参数的非接触式快速检测,提出了一种斜齿轮参数视觉检测方法。对于斜齿圆柱齿轮该方法可以实现齿顶圆直径、齿根圆直径、分度圆直径、齿数、模数、螺旋角和旋向的检测。通过预处理对斜齿轮端面画面进行图像增强,使用阀值分割和边缘检测获取斜齿轮端面轮廓,基于最小二乘法拟合齿顶齿根圆对斜齿轮的主要特征参数进行检测,在齿轮侧面使用SVM对斜齿轮的旋向进行分类检测,并计算齿顶区域的倾斜角度换算得螺旋角。经实验,该方法对不同规格斜齿轮的参数检测均有效,具有良好的稳定性、精确性。

In order to realize the non-contact rapid detection of helical gear parameters,a method of detecting helical gear parameters based on machine vision was proposed.For helical gears,this method can be used to detect the diameter of addendum circle,dedendum circle,reference circle,number of teeth,modulus,helix Angle and rotation direction.The image enhancement of the helical gear end face is carried out through preprocessing,then the threshold segmentation and edge detection are used to obtain the helical gear contour.After that,the contour is fitted based on the least square method to detect the main characteristic parameters ofthe helical gear.SVM was used to classify and detect the helical gear's spin direction on the side ofthe gear,and the helical Angle was converted by calculating the tilt Angle.The experimental results show that this method is effective in detecting the parameters of diferent helical gears with good stability and accuracy.