一种基于贝叶斯分类器的PCB焊点缺陷检测方法

A PCB Solder Spot Defect Detection Method Based on Bayesian Classifier

针对PCB板在焊接过程中存在多焊、少焊、漏焊等焊点表面质量问题,提出了一种基于朴素贝叶斯分类器的焊点缺陷检测方法。首先,通过图像分割,获取焊点ROI区域的表面特征,建立焊点的特征参数数据集;然后,将得到的特征参数根据其特征属性,计算出先验概率及条件概率,构建朴素贝叶斯分类器对数据集进行分类识别;最后,利用训练集和测试集对分类器分类结果进行实验验证。实验表明,该方法对焊点不良的检测率达到93.4%,能满足实际使用要求。

In order to solve the problems of welding spot surface quality in PCB welding process, an automatic welding spot surface defect detection method based on naive Bayesian classifier was proposed. Firstly, the surface characteristics of ROI region of solder spot were obtained by image segmentation, and the characteristic parameter data sets of solder dot were established. Then, the prior probability and conditional probability were calculated according to the characteristic attributes of the obtained characteristic parameters, and a naive Bayesian classifier was constructed to classify and recognize the data sets. Finally, training set and testing set were used to verify the classification results of the classifier. Experimental results showed that the detection rate of harmful welding spot could reach 93.4%, which could meet the practical requirements.