一种改进的印刷电路板缺陷检测分割算法

An Improved Segmentation Algorithm for Defect Detection of PCB

针对印刷电路(PCB)板缺陷检测分割中存在分割效果差、运行速度慢以及适用范围小的问题,提出一种改进的基于遗传算法的二维最大类间方差法的快速迭代算法。首先利用改进的遗传算法来确定分割中的最优阈值,再将这个最优阈值应用到二维最大类间方差法快速迭代算法中来确定最终的阈值最优解,从而完成分割。仿真实验表明,该算法分割的PCB图像,更加接近于人工标注的结果,最终的精度和Kappa系数达到了98.68%和0.970 6,具有广泛的应用前景。

A new and fast iterative algorithm for 2D-Otsu image segmentation method based on genetic algorithm was proposed,In order to solve the problems of poor segmentation effect,slow running speed and small scope were applied in the segmentation of PCB defect detection. First the improved genetic algorithm was used to determine the intermediate optimal threshold for segmentation. Then this optimal threshold is applied to the fast iterative algorithm for 2D-OTSU image segmentation method to determine the final optimal threshold value. Simulation experiments show that the segmentation of the PCB image,more close to the results of the manual labeling,the final accuracy and Kappa coefficient reached 98. 68% and 0. 970 6. And it has broad application foreground.