一种基于PCB板图像配准优化算法  

Research on PCB image registration optimization algorithm

作　　者：任永强[1] 吕华鑫 晏文彬 李润 REN Yongqiang;LYU Huaxin;YAN Wenbin;LI Run(School of Mechanical Engineering,Hefei University of Technology,Hefei 230009,China)

机构地区：[1]合肥工业大学机械工程学院,安徽合肥230009

出　　处：《合肥工业大学学报（自然科学版）》2022年第10期1305-1309,共5页Journal of Hefei University of Technology：Natural Science

基　　金：国家产业技术基础公共服务平台资助项目(2019-00899-2-1)。

摘　　要：为了提高PCB缺陷检测中的图像配准精度,文章提出一种结合梯度下降算法与随机抽样一致性(random sample consensus, RANSAC)算法的改进图像配准优化方法。对得到的灰度图像使用中值滤波去除噪声,通过拉普拉斯算子提取图像边缘来突出图像细节;使用尺度不变特征变换(scale invariant feature transform, SIFT)检测算法获取图像特征点并进行特征点匹配,通过匹配的特征点对之间的距离阈值来粗选出较强匹配点,使用改进的算法精选出强匹配点,同时算出基础图像变换矩阵;最后使用梯度下降法对基础图像变换矩阵进行拟合优化。实验结果表明,该算法在PCB板图像匹配过程中可以有效减少误匹配,并能得到准确的图像变换矩阵,且图像配准速度较快,能够满足实际工业现场检测要求。In order to improve the accuracy of image registration in PCB defect detection, this paper proposes an improved image registration optimization method combining gradient descent algorithm and random sample consensus(RANSAC) algorithm. Firstly, the median filter is used to remove the noise from the obtained gray image, and the edge of image is extracted by Laplace operator to highlight the details of image. Secondly, the scale invariant feature transform(SIFT) detection algorithm is used to obtain the image feature points and carry out feature points matching. Strong matching points are roughly selected by the distance threshold between the matched feature points, and the improved RANSAC algorithm is used to select the strong matching points. At the same time, the basic image transformation matrix is calculated. Finally, the gradient descent method is used to optimize the basic image transformation matrix. The experimental results show that the algorithm can effectively reduce mismatching in the process of PCB image matching and obtain the accurate image transformation matrix. The speed of image registration is fast which can meet the requirements of actual industrial field detection.

关 键 词：SIFT特征点 图像配准 变换矩阵 梯度下降 

分 类 号：TP391.413[自动化与计算机技术—计算机应用技术]