3D reconstruction of PCB based on ICP point cloud optimization

Abstract

Printed Circuit Board (PCB) was the basic carrier for a large number of components in electronic devices. The quality of PCB will directly affect the performance of electronic devices, so it is crucial to conduct defect detection during PCB assembly. Traditional defect detection methods are inefficient and costly, thus failing to satisfy the requirements of large-scale PCB inspection at this stage. Three-dimensional reconstruction has the advantages of non-contact and fast, and is gradually applied to PCB defect detection, but the general point cloud generation method is difficult to identify defects highly correlated with components, which is easy to produce missed detection. The Iterative Closet Point (ICP) point cloud alignment optimization method proposed in this paper can effectively identify such height defects. Firstly, the point cloud is matched in a two-by-two alignment, and the third piece is aligned on the basis of the completion of the first two pieces of matching, and so on. Then the duplicate points are removed to ensure the 3D morphological features of the point cloud while effectively filtering out the redundant point cloud data points, and finally the non-overlapping parts are superimposed to optimize the point cloud model. The experimental results show that the average alignment error of the algorithm in this paper is less than 0.05% on the PCB board, which can be better applied to detect defects in the PCB assembly process.