Design of gear defect detection system based on machine vision

As the society develop rapidly, the existence of machinery is indispensable, inside the most basic part is the gear. But the traditional measurement methods such as vernier calipers, coordinate measuring instruments and so on, greatly reduce the measurement efficiency under the condition of ensuring accuracy. In view of the above problems, this paper designs a gear defect detection system based on machine vision. The hardware part of the system uses a white ring light source, a area array camera model with ME2L-505-36U3M/C and a prime focus lens. In the software part, Halcon software building the gear parameter detection system combining with DobotStudio platform is used to control the robot arm. Firstly, the reality coordinates and pixel coordinate are converted into matrix through nine-point calibration. Secondly, images are filtered by Gaussian smoothing, and detect edge through Canny operator. At that way, we can obtain the basic parameters of the gear quickly and accurately, and detect whether the gear qualified. Thirdly, the corresponding coordinate will be transferred to DobotStudio platform via the TCP transmission protocol, so that the robot arm will be controlled to complete the screening work. Through experimental debugging, the measurement parameter error of the system is less than 0.10mm. Unqualified gears are absorbed by the mechanical arm with the air pump and placed aside, so as to achieve the purpose of gear screening. The experimental results show that the system can greatly improve the efficiency and accuracy of gear defect detection, and has certain application value.