基于图像识别的光纤陶瓷插芯表面裂痕检测装置设计与实现

Design and Implementation of Surface Crack Detection Device for Fiber Optic Ferrule Based on Image Recognition

针对光纤陶瓷插芯次品中有裂痕的问题,研制出一种基于图像识别的光纤陶瓷插芯表面裂痕检测装置,能够在节省人工成本的同时达到高灵敏度和高健壮性。首先介绍这种检测装置的系统硬件设计和软件设计,描述图像处理识别光纤陶瓷插芯裂痕的基本方法;其次用工业相机对光纤陶瓷插芯表面进行图像实时采集,运用图像处理识别算法进行最优化设计;最后通过实验来验证该裂痕检测识别方法的可行性。实验结果表明:工业相机可以清晰采集光纤陶瓷插芯表面的裂痕,通过把采集的图像实时传送至计算机,计算机运用图像处理识别算法可以准确检测出光纤陶瓷插芯表面的裂痕。

In order to solve the problem of cracks in the defective fiber optic ferrule,a surface crack detection device for fiber optic ferrule based on image recognition is developed,which can achieve high sensitivity and robustness while saving labor costs.Firstly,the system hardware and software design of this detection device are introduced,and the basic method of image processing for identifying cracks in fiber optic ferrule is described.Secondly,an industrial camera is used to collect real-time images of the surface of the fiber optic ferrule,and image processing recognition algorithms are used for optimization design.Finally,the feasibility of the crack detection and recognition method is verified through experiments.The experimental results show that industrial cameras can clearly capture cracks on the surface of fiber optic ferrule.By transmitting the collected images in real-time to a computer,the computer measurement can accurately detect cracks on the surface of fiber optic ferrule using image processing recognition algorithms.