Transforming PCB Solder Joint Detection with Deep Learning Empowered X-ray Nondestructive Testing

In order to avoid failures and quality problems caused by welding defects, PCB solder joint inspection has become a crucial step in the production and manufacturing of electronic products. Using deep learning-based digital radiographic nondestructive inspection for inspecting solder joints inside PCB boards can alleviate labor pressure and enhance productivity. Based on RetinaNet, three datasets containing common digital radiographic images of PCB solder joint defects were established, and an automatic detection network model was built. After training and testing, the model achieved an average detection accuracy of 92.7% for defective images. Experimental results demonstrate that the proposed model effectively enhances the performance and efficiency of PCB solder joint defect detection under X-ray.