| Title |
Optimal UV Light Intensity Selection Algorithm for Detecting the Tiny Bubble in Conformal Coating |
| DOI |
https://doi.org/10.5573/ieie.2024.61.5.40 |
| Keywords |
Conformal coating; Bubble detection; UV light intensity; Feature space; YOLO v8 |
| Abstract |
Conformal coating is a key technology used to protect Printed Circuit Boards (PCBs). And the bubble inspection is an essential task for the reliability of the PCB since the bubbles which generated during the coating process can be causes of circuit malfunctions. The visibility of bubbles in images captured under UV (Ultra Violet) light for bubble inspection is influenced by various factors, including the background color of the PCB, the thickness of the coating as well as UV light intensity. This paper proposes a technique for selecting the optimal UV intensity level for bubble detection when the information about the coating thickness is not available. In the proposed method, the effective region is extracted from the input image, and then two features, that is average brightness of B component and maximum position of H component, are calculated. And the optimal UV light intensity is determined based on the Mahalanobis distance considering the statistical characteristics of the data in the feature space. Meanwhile the YOLO v8 model is adopted in the bubble detection stage. The bubbles are classified into multi according to location to consider the irregularity of bubble shapes and the image tiling is applied to enhance the detection performance for tiny bubbles. The validity of the proposed algorithm is demonstrated through experiments which investigate the correlation between the bubble detection rate and Mahalanobis distance according to UV intensity level. It is expected to provide a more reliable bubble inspection solution by integrating the proposed algorithm into the existing conformal coating equipment. |