中文摘要：

提出了基于齿轮外轮廓最小凸包及凸包缺陷的直齿圆柱齿轮齿数及缺齿识别算法,采用二值图像连通域跟踪提取齿轮外轮廓,然后进行外轮廓最小凸包及凸包缺陷检测,去除伪凸包缺陷后进行齿数检测及缺齿部位识别。根据直齿圆柱齿轮外形特征,提出了基于齿轮外轮廓凸包顶点的齿顶圆检测方法,以及基于齿轮外轮廓凸包缺陷最深点的齿根圆拟合方法,实现了直齿圆柱齿轮的特征圆参数快速测量。实验结果表明,齿数及缺齿识别算法正确率高达98.48%;与现有两种典型方法对比,齿轮特征圆拟合算法可以实现各种情况下齿顶圆和齿根圆的高速高精度定位。

英文摘要：

A spur gear tooth number and missing teeth identification algorithm based on minimum convex hull and convex defects of outer contour is presented in the paper. Firstly,the connected domain of binary image is adopted to trace the outer contour of gear. Then,the minimum convex hull and convex defects of outline are detected,and fake convex defects need to be removed before identifying tooth number and missing teeth part. According to the shape characteristics of spur gear,an addendum circle detection method based on the convex hull vertices of gear outer contour and a dedendum circle fitting method based on the deepest point of convex defects are proposed,which have realized the rapid measurement of characteristic circle parameters of spur gear. The experimental results show that the accuracy of the proposed tooth number and missing teeth recognition algorithm is as high as 98. 48%. Compared with the existing two kinds of typical methods,the proposed characteristic gear circle fitting method can achieve high speed and high precision positioning of addendum circle and dedendum circle under various conditions.