Classification of Spot-Welded Joints in Laser Thermography Data Using Convolutional Neural Networks

Spot welding is a crucial process step in various industries. However, classification of spot quality still tedious due to the complexity and sensitivity test material, which drain conventional approaches its limits. In this article, we propose an approach for inspection weldings using images from laser thermography data. We data preparation based on underlying physics spot-welded joints, heated with pulsed by analyzing intensity over time derive dedicated filters generate training datasets. Subsequently, utilize convolutional neural networks classify weld compare performance different models against each other. achieve competitive results terms classifying classes compared traditional approaches, reaching accuracy more than 95 percent. Finally, explore effect augmentation methods.