基于机器视觉的卡簧装配缺失检测

Detection of Circlip Missing Assembly Based on Machine Vision

针对在循环检测过程中不同的工件表面特征易出现变化而难以检测的问题,提出了一种基于机器视觉的游艇换挡执行机构卡簧装配缺失检测方法。通过对采集的图像进行预处理、边缘检测、轮廓提取、轮廓拟合定位、掩模屏蔽、二值化及形态学滤波等操作,将图像中卡簧特征未发生明显变化的区域分割出来,利用该特征区域像素数量来检测卡簧的装配完整性。对轮廓的拟合定位部分和传统的霍夫圆检测定位方法做了对比,结果表明,利用传统的霍夫圆检测定位方法平均运行时间为13.99ms,但定位准确率只有87.33%,而基于机器视觉的检测方法平均运行耗时为12.56ms,定位准确率达到98.7%,因而能够实现较好的检测效果。

Aimed at the problem that the surface features of different workpieces were easy to change during the cycle detection process that leaded to difficult detection,a method for detecting the missing assembly of the yacht shift actuator was proposed.After pre-processing,edge detection,contour extraction,contour fitting positioning,mask masking,binarization and morphological filtering of the acquired images,the regions in the image in which the circlip features had not changed significantly were segmented.The number of pixels of the feature area was utilized to detect the assembly integrity of the circlip.The fitting and positioning part of the contour was compared with the traditional Hough circle detection and positioning method.The experimental results showed that the average running time of the traditional Hough circle detection and positioning method was 13.99 ms,but the positioning accuracy was only 87.33%.The average running time of the method was 16.56 ms,and the positioning accuracy was 98.7%.Therefore,the better detection effect can be achieved.