基于机器视觉的护套编织线缺陷检测系统设计

针对导线编织护套在编织过程中存在的几种缺陷问题进行分析.利用滤波的方式获得清晰图片,由于采集过程为动态过程,需对图像坐标进行校对,并结合模板的基准坐标进行图像分割.针对编织线表面存在的黑斑、色差等缺陷问题制定模板图像,通过对实时检测的图像与模板图像进行比对运算得到差异图像,再对差异图像进行二值化处理和滤波得到差异特征,差异特征的区域和大小与设定的差异阈值比对,来判定实时图像是否存在缺陷.针对编织过程中经常出现的毛球缺陷进行分析,利用边缘检测、二值化处理和模板比较的方式得到边缘外的信息,根据其形状及面积大小与设定阈值的比对判定是否存在缺陷.经过对实时图像进行多次试验,其结果均能够准确地判断出编织线的缺陷问题.

Detection System Design for Sheath Braid Defect Based on Machine Vision

Several defects of braided sheath during the braiding process of superconducting wire are analyzed.First,a clear picture is obtained by filtering,as the image acquisition is a dynamic process,the image coordinate needs to be corrected,and the image is segmented by the reference coordinates of the template.To eliminate the defects of freckle and color difference on the surface of braided wire,the template image is formulated.By comparing the real-time detection image and the template image,the difference image is obtained,and the difference feature is obtained by binary processing and filtering,thus to determine whether the real-time image is defective by comparing the region and size of the difference feature and the set differential threshold.Secondly,by analyzing the clews defects in braiding process,the information out of edge is obtained by edge detection,binary processing and template comparison,thus to identify a defect by comparing the shape and area size and the thresholds.Through many experiments on real-time images,the results can accurately determine the defects of the braided wire.