基于机器视觉的齿轮接触疲劳点蚀实时检测方法

Real-time Detection Method of Gear Contact Fatigue Pitting Based on Machine Vision

为了提高齿轮接触疲劳试验中试样疲劳点蚀的检测精度和检测效率,实现齿轮点蚀的可视化、量化以及实时检测,提出了基于机器视觉的齿轮接触疲劳点蚀实时检测方法。基于齿轮啮合原理和线阵相机拍摄原理,建立齿轮接触疲劳试验检测系统,得到拍摄齿轮齿面的最佳偏心拍摄距离,提出了齿轮啮合面图像修正算法,并采用疲劳特征初步检测和精确检测相结合的策略,提高了疲劳点蚀的检测精度。试验结果表明,采用所提出的局部阈值测量方法对齿轮接触疲劳点蚀进行检测时,点蚀检测的平均绝对误差为0.12160 mm^(2),平均相对误差为2.2188%,检测精度达到了97.7812%。所搭建点蚀检测系统具有可视化、可量化、效率高等优点,是一种新的齿轮接触疲劳点蚀检测平台。

To improve detection accuracy and efficiency of sample fatigue pitting in gear contact fatigue test,and to realize the visualization,quantification and real-time detection of gear pitting,real-time detection method of gear contact fatigue pitting based on machine vision is proposed.Based on principle of gear meshing and shooting principle of line array camera,gear contact fatigue test detection system is established,and the optimal centrifugal shooting distance of gear flank is obtained by analyzing gear rotation process.The image overlap phenomenon caused by inconsistency of each point speed and camera set line frequency on gear profile is proposed,and image correction algorithm of gear meshing surface is proposed,which improves accuracy of gear contact fatigue pitting detection results.The combination of preliminary detection and precision detection of fatigue characteristics is used to improve detection accuracy of fatigue pitting.The experimental results show that average absolute error of pitting detection is 0.12160 mm^(2),average relative error is 2.2188%,and detection accuracy is 97.7812%.The pitting detection system has advantages of visualization,quantification,high efficiency,real-time monitoring and failure judgment,which can provide a new experimental method for gear contact fatigue pitting detection.