基于机器视觉的紧固套缺陷检测方法估计

Defect Detection Method of Fastening Sleeve Based on Machine Vision

针对紧固套缺陷检测中人工检测效率低等问题,设计了基于机器视觉的紧固套缺陷检测方法。首先根据紧固套的特点搭建了图像采集平台,突出缺陷特征;然后基于Halcon设计了相应的缺陷检测方法:先通过模板匹配的方法对检测区域定位;再使用基于频域图像差分的方法抑制干扰,突显缺陷特征;然后使用改进的动态阈值法得到缺陷备选图;最后使用面积以及长径两个特征提取出缺陷。实验结果表明:该方法在1500张图像上的测试准确率为98.5%,漏检率为0,过检率为1.4%,可以达到实际工业生产的要求。

Aiming at the problem of low manual inspection efficiency in fastening sleeve defect detection,a machine vision-based fastening sleeve defect detection method is designed.First,an image acquisition platform was built according to the characteristics of the fastening sleeve to highlight the defect features;then a corresponding defect detection method was designed based on Halcon:firstly locate the detection area through the template matching method;then use the method based on frequency domain image difference to suppress interference,highlight the defect feature;next use the improved dynamic threshold method to get the defect candidate map;finally use the two features of area and long diameter to extract the defect.The experimental results show that the test accuracy rate of the method on 1500 images is 98.5%,the missed detection rate is 0,and the over-detection rate is 1.4%,which can meet the requirements of actual industrial production.