基于机器视觉的纽扣缺陷检测算法研究

Research on Button Defect Detection Algorithm Based on Machine Vision

针对传统人工检测纽扣效率低,成本高,外界因素影响大等问题,提出基于机器视觉的纽扣缺陷检测算法。通过均值滤波平滑图像,以减少噪声点;再用Canny算子获取图像边缘特征,判断是否存在大小眼或者扣眼堵塞;最后采用Otsu分块阈值的方法进行疵点提取。实验结果表明,此方法检测缺陷的准确率高达97%,且效率高,有良好的工业实用价值。

In view of the problems of low efficiency,high cost and great influence of external factors in traditional manual detection of buttons,an algorithm of button defect detection based on machine vision is proposed. Smoothing the image by means ofthe mean filter to reduce the noise points; Then the Canny operator is used to obtain the edge features of the image to determinewhether there are large or small eyes or obliterated eyes. Finally,Otsu block threshold is used to extract defects. The experimentalresults show that the accuracy of this method is as high as 97%,with high efficiency and good practical value in industry.