Total: 1
This paper presents a novel approach for detecting soldering defects in Printed Circuit Boards (PCBs) composed mainly of Surface Mount Technology (SMT) components, using advanced computer vision and deep learning techniques. The main challenge addressed is the detection of soldering defects in new components for which only examples of good soldering are available at the model training phase. To meet industrial quality standards, we must keep the leakage rate (i.e., miss detection rate) low. To address this, we design the system to be "unknown-aware" with a low unknown rate and utilize the knowledge gained from the soldering examples of old components to detect the soldering defects of new components. We evaluated the method on a real-world dataset from an electronics company. It significantly reduces the leakage rate from 1.827\% $\pm$ 3.063\% to 0.063\% $\pm$ 0.075\% with an unknown rate of 3.706\% $\pm$ 2.270\%, compared to the baseline approach.