Research On Optical Fiber Quality Inspection Based On Machine Vision

Optical fiber has been more and more widely used in communication field,industrial measurement field and medical field,etc.,because of its efficient and rapid information transmission characteristics,stable chemical performance as well as its super anti-electromagnetic and radiation interference performance and stable chemical performance.People also put forward higher and higher requirements for the quality of optical fiber.In the medical field,the quality of optical fiber may greatly influence the objective conditions of surgery.Therefore,in the production stage of optical fiber,it is of great significance to measure and detect the quality parameters of optical fiber rapidly and efficiently.This paper carried out the research on the intelligent platform of optical fiber quality parameter detection based on machine vision.The main contents of the research are as follows:(1)According to the requirements of the project,build a platform for optical fiber quality detection.The platform mainly includes stable laser source,optical fiber clamp,three-dimensional electric translation table and two-dimensional electric angle table,high-precision industrial camera,self-designed optical fiber collimation adjustment controller and upper computer software.(2)Designed optical fiber collimation adjustment controller.The main functions of the controller are as follows: control the five stepper motors located on the twodimensional electric angle table and the three-dimensional electric translation table to adjust the position and incidence angle of the fiber to be measured;communication with upper computer through serial port module to coordinate the collimation control of the fiber beam;the input and display modules of the collimator control regulator are designed to facilitate subsequent function expansion.In this paper,we have finished the design of optical fiber collimating controller and debugging,and related module driver and control software.(3)The classical edge detection method based on differential is used to extract the light spot contour.The center location algorithm is improved by combining the least square fitting method and the weighted center of gravity method.The least square fitting method was used for ellipse fitting,and then the ellipse fitting was optimized based on the Lester criterion to improve the precision of beam width extraction.According to the actual situation,the collimation scheme of optical path is developed.(4)Finished the development of upper computer interaction interface based on MFC(Microsoft Foundation Classes)framework;According to the requirement,the communication mode between the upper computer and the fiber collimating control regulator is designed,and the communication protocol is defined.C++ language is used to realize the algorithms of spot image denoising,spot contour extraction,center positioning and spot beam width extraction.According to the theoretical of optical fiber,the calculation of optical fiber parameters is completed.(5)The test and error analysis are carried out on the optical fiber quality detection intelligent platform based on machine vision.The experimental data show that the optical fiber quality detection intelligent platform designed in this paper has completed all the design indexes,and the detection error is less than 2%,meeting the design requirements.