Cob Angle Recognition Algorithm Based On Machine Vision

Abstract

In order to solve the problem of lens skew when some products of single channel cob coupling machine are coupled in actual production, the recognition algorithm of lens and PCB in cob optical module based on machine vision is studied. Industrial camera is used to collect lens and PCB image. Halcon software is used to preprocess lens and PCB image and detect edge to identify the target. Finally this papaer calculates the angle difference between lens and PCB. Then, through the platform verification, the results show that: the recognition of lens and PCB and the calculation of angle difference can be realized in 0.5 seconds, and the recognition effect can still be guaranteed when there are miscellaneous points in the background, and the repetition accuracy can reach 0.05 ° and the error between the measured value and the actual value is within 0.2 ° which meets the requirements of industrial production. This research has a certain reference significance for visual recognition in the actual production of cob optical module.