基于改进亚像素边缘检测的气门弹簧节距测量  

Valve Spring Pitch Measurement Based on Improved Sub-pixel Edge Detection

作　　者：王虎 陈鸣伦 WANG Hu;CHEN Minglun(School of Mechanical Engineering,Hefei University of Technology,Hefei 230041,China)

机构地区：[1]合肥工业大学机械工程学院,合肥230041

出　　处：《机械工程师》2024年第7期15-18,共4页Mechanical Engineer

摘　　要：针对传统弹簧节距测量效率低、精度差等问题,提出一种基于高斯函数模型的亚像素弹簧边缘检测方法。首先对图像进行Canny边缘检测定位,然后寻找梯度方向,进而采用高斯函数对梯度方向上的灰度梯度幅值进行拟合,按照定步长的平均值法来确定亚像素边缘点的亚像素坐标,并对亚像素坐标采用最小二乘法拟合圆,根据圆心坐标计算出弹簧节距。经测试,该方法能快速测量出弹簧节距,将平均计算时间降低到传统拟合亚像素检测方法的21%。To solve the problems of inefficiency and low precision of traditional spring pitch measurement,this paper proposes a sub-pixel spring edge detection method based on Gaussian function model.First,Canny algorithm is used to detect the edge of image for ensuring the edge location and find the gradient direction.Then gaussian function is made to fit the gradient magnitude of gray along the gradient,which is based on the fixed step average method to determine the coordinates of sub-pixel edge points and use the least squares fitting circle according to the results.Finally,spring pitches are calculated by center coordinates.Through the test,this method can quickly measure the spring pitch and reduce the average calculation time to 21%of the traditional fitting sub-pixel detection method.

关 键 词：机器视觉 边缘检测 图像处理 气门弹簧 亚像素算法 

分 类 号：TP391.7[自动化与计算机技术—计算机应用技术]