Research Of Optical Fiber Ferrules Concentricity Detection Based On Machine Vision

Optical fiber connector is the essential passive device which has largest usage in optical fiber communication system.Its core component is optical fiber ferrules.The optical fiber ferrules are made of ceramic,so they also known as ceramic ferrules,ceramic ferrules are used for precision alignment when optical fiber is docking.Its concentricity directly affects the insertion loss and reflection loss of optical fiber connector,and then affects the transmission performance of the entire optical fiber communication system,so the concentricity of ceramic ferrules is an important parameter to judge the quality of optical fiber connectors.The detection for the concentricity of ceramic ferrules,the accuracy of the main detection methods currently is not high with different degree,and some methods are manual detection.Detection method based on machine vision has the advantages of high accuracy,fast speed,no contact,no damage and so on,and it has been widely used in the filed of industrial detection.After investigation and research at ceramic ferrules manufacture factory,detection method based on machine vision applied in the detection for the concentricity of ceramic ferrules has a broad market prospect.So this paper researches a algorithm based on machine vision for concentricity detection of ceramic ferrules,and designs the prototype of the concentricity detection system,it is in the trial stage currently.In this paper,many documents which about the concentricity detection of ceramic ferrules in the domestic and overseas are consulted.On the basis of reference to the working principle of the similar products in the ceramic ferrules manufacturer factory,this paper combines production techniques and workshop environment,and then it carried out in-depth research and repeated testing from the system structure design,software interface design and algorithm design.Firstly,according to the characteristic of material and structure of ceramic ferrules,selects the light source,illumination mode,lens and camera,and designs system structure of instrument according to the optical path structure;then designs the system software and interface in the VS2010 development platform according to the functionalrequirements for detection system by the manufacture factory;then according to the characteristics of images which collected from the end face of ceramic ferrules,detects edge at pixel level,then designs a sub-pixel edge detection algorithm based on Gauss function fitting,gets the inner circle center coordinates by fitting with the detected sub-pixel edge of inner hole of ceramic ferrules;finally designs a algorithm for calculating concentricity by fitting trajectories of inner holes' center in the process of rotation,the diameter of the circle which got by fitting is the concentricity of ceramic ferrules.A series of experiments have been designed in this paper,they compare the detection results of ceramic ferrules in the same batch with the same type of product,verify the accuracy of the algorithm and the detection system which designed in this paper;and they analyze the detection results of ceramic ferrules in different batch production,verify the stability of the algorithm and the detection system which designed in this paper.