Based on the Deep Study of 3D Printing Defect Detection Technology Research

With the rapid development of 3D printing technology, surface defect detection, as one of the important methods for its quality assurance, has attracted more and more attention. In the process of 3D printing, the defect detection of the printing layer on the surface of the part can find the appearance quality problems in the printing in time, and avoid the subsequent serious metallurgical defect quality problems. In this paper, the surface point cloud data of printed parts obtained by 3D vision is combined with deep learning technology to realize the defect detection of 3D printed parts. The improved methods include replacing the original convolution with the residual network structure, embedding the attention mechanism module combining channels and Spaces, and improving and optimizing U-Net by using the weighted cross-entropy loss function and Adam optimization algorithm according to the defect proportion. Through multiple sets of comparison experiments, the results show that the improved U-Net defect segmentation algorithm can effectively segment the defect area on the two-dimensional depth map, and classify the defect, so as to realize the defect detection and recognition of the surface of 3D printed parts.