基于NCC的PCB图像配准算法研究

Research on PCB Image Registration Algorithm Based on NCC

为解决PCB数码喷印中图像数据量大导致传统PCB图像配准时间复杂度高的问题,研究归一化互相关匹配算法(NCC),从候选点集和特征块选取两个角度对算法效率优化。利用统计滤波的方法滤除候选点中的误差点,减少计算量;分析PCB图像中配准点特点,利用动态规划的思路,并基于广度优先搜索方法,对特征块自适应选取,从而减少配准点集的遍历。实验中,利用分割后的基板图像与Gerber图像配准建立空间位置关系,再使用改进的配准算法来配准。实验表明,配准精度在99%左右,效率提高至1.6倍。

In PCB digital printing,the large amount of image data leads to the high complexity of traditional PCB image registration,which is the main factor that affects the efficiency of the system.By studying the image registration algorithm,based on the normalized cross correlation matching algorithm(NCC),the efficiency of the algorithm is optimized from two perspectives of candidate point set and feature block selection.Statistical filtering is used to filter out the error points in the candidate points and reduce the amount of calculation;the characteristics of the registration points in the PCB image are analyzed.By using the idea of dynamic programming,and based on the breadth first search method,feature blocks are adaptively selected,thereby reducing the registration points Set traversal.Experiments show that this method can effectively improve the traditional normalized cross correlation registration method,that the registration accuracy is about 99%,and that the efficiency is increased by 1.6 times.