||During the image inspection of RDL products, PCB manufacturers encountered the problem of confusion between circuit layers due to the use of transparent substrates (dielectric materials), which prevented automatic inspection of the current manufacturing process. The research purpose of this article is to develop an image capture system that filters out the interference of the lower layer circuit due to the transparent substrate in the RDL product, and presents a high contrast between the upper layer circuit and the background.|
After a series of tests, this study successfully developed the "enhanced RDL circuit inspection system using fluorescence microscopy technology" through the fluorescent reaction of the substrate in the RDL structure to improve the failure of AOI in RDL circuit inspection. question. In the experiment, using the characteristics of the fluorescent reaction of the substrate, the original detection system based on the contrast between the color difference between the circuit and the background in the image was replaced by the difference in fluorescence between the two in the image, and short-wavelength light was used in the transparent substrate. The characteristic of fast attenuation avoids confusion between circuit layers and layers. With regard to the improvement of the detected image, the contrast between the upper circuit and the background has increased from 5% to 10% to over 80%, and the pre-process image which mainly caused interference has also been reduced from 47% to less than 20%. It greatly reduces the difficulty of detecting upper-layer line extraction, making RDL line detection possible.
The fluorescent image capture system developed in this research can provide the most important short circuit and open circuit detection on the line, and its feasibility has been confirmed through actual sample tests. In terms of some specific RDL circuit detection, this study also proposes a dual-image image capture method based on the research and construction experience of the fluorescent microscope. This method is currently only in the design stage, and has not yet completed the structure and feasibility verification.
|| J. H. Lau, “Fan-Out Wafer-Level Packaging,” Springer (2018).|
 N. Hayao, “NTI 100: The Unsinkable, Unstoppable PCB Market.” (2022).
 林定皓，「電路板電氣測試與AOI檢驗技術簡介.」，臺灣電路板協會 (2008).
 王家鴻，「Approach of a Reliable Solder Bump with RDL Structure for WLCSP Application Study.」，國立交通大學，碩士論文，民國99年.
 G. R. Fowles, "Introduction to Modern optics," 2nd edition, Dover (1968).
 J. Kapuscinski, “DAPI: a DNA-specific fluorescent probe.” Biotech Histochem, 70:220–233 (1995).
 H. Yasuhiko, D. Hideaki, K. Koichi, and I. Tadashi, “A System for PCB Automated Inspection Using Fluorescent Light,” IEEE Transactions on Pattern Analysis and Machine Intelligence, 69–78 (1988).
 J. F. Waymouth, “Handbook of Advanced Lighting Technology,” Springer (2017).
 J. Lichtman, J. A. Conchello, “Fluorescence Microscopy,” Nature Methods, 2:910–919, (2006).
 Joseph R. Lakowicz, “Principles of Fluorescence Spectroscopy,” Springer (2006).
 A. A. Michelson, “Studies in Optics” U. Chicago Press (1927).
 R. C. Gonzalez, R. E. Woods, “Digital Image Processing,” 4th edition, Pearson (2018).
 Y. Pang, M. Sun, X. Jiang, X. Li, "Convolution in Convolution for Network in Network," in IEEE Transactions on Neural Networks and Learning Systems, vol. 29, no. 5, pp. 1587-1597 (2018).
 E. S. Gedraite, M. Hadad, “Investigation on the effect of a Gaussian Blur in image filtering and segmentation,” Proceedings ELMAR-2011, pp. 393-396 (2011).
 R. M. Haralick, S. R. Sternberg, X. Zhuang, "Image Analysis Using Mathematical Morphology," in IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. PAMI-9, no. 4, pp. 532-550 (1987).
 Daniel Malacara, “Geometrical and Instrumental Optics,” Academic Press (1988).
 林良平, “光學顯微鏡之最新進展－無限遠補正光學系統. ” 科儀新知第二十二卷第四期 90.2
 顯微物鏡(Edmund Optics).
 Bill Lubanovic 著，賴義民 譯，「精通Python」，基峯 (2020).
 李立宗，“科班出身的 AI人必修課：OpenCV 影像處理 使用 Python,” ，深智數位 (2019).