基于PSO-SVM的FPC焊盘表面缺陷检测研究

Study on Defect Detection of Flexible Printed Circuit Solder Surface Based on PSO-SVM

针对柔性线路板(FPC)焊盘表面的缺陷检测,建立了一种利用粒子群算法(PSO)进行参数寻优的PSO-SVM分类识别模型。首先通过OTSU法将焊盘从原始图像中分割出来,然后对其5种表面缺陷从形状、灰度、纹理三个方面提取了14维特征,接着用粒子群算法方法对支持向量机的参数优化以获得较高的识别准确率,最后对缺陷样本进行分类识别,并将其与GS-SVM和BP神经网络分类性能进行对比。实验证明了该方法可以对焊盘缺陷进行准确的分类识别。

Aiming at the surface defect detection of flexible printed circuit(FPC)solder,a PSO-SVM classification recognition model based on particle swarm optimization(PSO)for parameter optimization is established.The model firstly separates the solder from the original image by morphology and OTSU method,and then extracts 14-dimensional features from three aspects of shape,gray scale and texture for five surface defects,Then,particle swarm optimization was used to optimize the parameters of SVM to obtain high recognition accuracy.Finally,defect samples were classified and identified and compared with the classification performance of GS-SVM and BP neural network.Experimental results show that this method can be used to classify and identify the defects accurately.