自2020年起，新冠病毒肺炎（COVID-19）疫情推動了科技業員工大規模轉向居家辦公，這一轉變結合工業4.0時代的推進，顯著促進了對個人電腦及其周邊硬體、電動車等技術的需求增長，進一步推動了電子消費市場對半導體的需求。在此背景下，發光二極體（light-emitting diode, LED）因其節能特性、小體積和高亮度，成為一種關鍵的電子產品。然而，在LED生產過程中，晶片內部空洞（VOID）缺陷的檢測，目前仍主要依賴於傳統的人工目視檢查。生產線上的人工檢測方式不僅耗時、標準不一致，而且識別效率低，導致異常分類準確率低。
LED缺陷檢測中遭遇影像辨識準確率低下問題以及後續資料整合分散的痛點，本研究提出了一種創新的HITMDI（Hybrid Image Training Move Data Integration）系統架構，此方法結合傳統影像處理技術與YOLO神經網路方法，透過實驗驗證評估指標。YOLOv9結合電腦視覺之混和模型的平均精確約為91.13%。儘管深度學習在許多方面展示其明顯優勢，但持續改進和優化傳統計算機視覺技術中的演算法有潛力使其精確度超越基於深度學習的架構。
因此，結合兩種方法的優點可提供更加強大且適應性更強的影像辨識解決策略方案。綜合本研究範圍，電腦視覺技術主要涉及圖預處理以及提取每張異常的標準樣品圖像，而深度學習的部分則著重探討YOLOv9結合電腦視覺混和訓練。本研究不僅豐富了現有影像辨識技術的理解，同時也指出了未來可能的發展方向，包括更有效地結合電腦視覺技術與先進的深度學習方法來滿足特定的工業技術需求。

Since 2020, the COVID-19 epidemic has driven a large-scale shift of employees in the technology industry to working from home. This shift, combined with the advancement of the Industry 4.0 era, has significantly promoted the demand for personal computers and peripheral hardware, electric vehicles, etc. The growth in demand for technology has further promoted the demand for semiconductors in the consumer electronics market. In this context, light-emitting diodes (LEDs) have become a key electronic product due to their energy-saving properties, small size and high brightness. However, in the LED production process, the detection of voids inside the wafer (VOID) defects still mainly relies on traditional manual visual inspection. The manual detection method on the production line is not only time-consuming and inconsistent with standards, but also has low identification efficiency, resulting in low anomaly classification accuracy.
However, In LED defect detection, there are challenges such as low image recognition accuracy and scattered data integration. This study proposes an innovative HITMDI (Hybrid Image Training Move Data Integration) system architecture, which combines traditional image processing techniques with YOLO neural network methods to validate evaluation metrics through experiments. The average accuracy of the YOLOv9 combined with computer vision hybrid model is about 91.13%. Although deep learning has demonstrated its obvious advantages in many aspects, continuous improvement and optimization of algorithms in traditional computer vision technology have the potential to surpass deep learning based architectures in terms of accuracy.
Therefore, combining the advantages of both methods can provide a more robust and adaptable image recognition solution. Based on the scope of this study, computer vision technology mainly involves image preprocessing and extracting standard sample images for each anomaly, while the deep learning part focuses on exploring YOLOv9 combined with computer vision blended training. This study not only enriches the understanding of existing image recognition technologies, but also points out possible future development directions, including more effectively combining computer vision technology with advanced deep learning methods to meet specific industrial technology needs.

6參考文獻
[1] M.-H. Fang, Z. Bao, W.-T. Huang, and R.-S. Liu, "Evolutionary generation of phosphor materials and their progress in future applications for light-emitting diodes," Chemical Reviews, vol. 122, no. 13, pp. 11474-11513, 2022.
[2] H. E. Martz, C. M. Logan, D. J. Schneberk, and P. J. Shull, X-ray Imaging: fundamentals, industrial techniques and applications. CRC Press, 2016.
[3] A. Du Plessis, I. Yadroitsev, I. Yadroitsava, and S. G. Le Roux, "X-ray microcomputed tomography in additive manufacturing: a review of the current technology and applications," 3D Printing and Additive Manufacturing, vol. 5, no. 3, pp. 227-247, 2018.
[4] K.-J. Choi, Y.-H. Lee, J.-W. Moon, C.-K. Park, and F. Harashima, "Development of an automatic stencil inspection system using modified Hough transform and fuzzy logic," IEEE Transactions on Industrial Electronics, vol. 54, no. 1, pp. 604-611, 2007.
[5] X. Peng, Y. Chen, W. Yu, Z. Zhou, and G. Sun, "An online defects inspection method for float glass fabrication based on machine vision," The International Journal of Advanced Manufacturing Technology, vol. 39, pp. 1180-1189, 2008.
[6] D. Chetverikov and A. Hanbury, "Finding defects in texture using regularity and local orientation," Pattern Recognition, vol. 35, no. 10, pp. 2165-2180, 2002.
[7] D. G. Lowe, "Object recognition from local scale-invariant features," in Proceedings of the seventh IEEE international conference on computer vision, 1999, vol. 2: Ieee, pp. 1150-1157.
[8] T. Ojala, M. Pietikainen, and T. Maenpaa, "Multiresolution gray-scale and rotation invariant texture classification with local binary patterns," IEEE Transactions on pattern analysis and machine intelligence, vol. 24, no. 7, pp. 971-987, 2002.
[9] S.-H. Chen and D.-B. Perng, "Automatic optical inspection system for IC molding surface," Journal of Intelligent Manufacturing, vol. 27, pp. 915-926, 2016.
[10] M. Ferguson, R. Ak, Y.-T. T. Lee, and K. H. Law, "Automatic localization of casting defects with convolutional neural networks," in 2017 IEEE international conference on big data (big data), 2017: IEEE, pp. 1726-1735.
[11] Y. He, K. Song, Q. Meng, and Y. Yan, "An end-to-end steel surface defect detection approach via fusing multiple hierarchical features," IEEE transactions on instrumentation and measurement, vol. 69, no. 4, pp. 1493-1504, 2019.
[12] S. Tang, F. He, X. Huang, and J. Yang, "Online PCB defect detector on a new PCB defect dataset," arXiv preprint arXiv:1902.06197, 2019.
[13] W. Du, H. Shen, J. Fu, G. Zhang, X. Shi, and Q. He, "Automated detection of defects with low semantic information in X-ray images based on deep learning," Journal of Intelligent Manufacturing, vol. 32, pp. 141-156, 2021.
[14] T.-Y. Lin, P. Dollár, R. Girshick, K. He, B. Hariharan, and S. Belongie, "Feature pyramid networks for object detection," in Proceedings of the IEEE conference on computer vision and pattern recognition, 2017, pp. 2117-2125.
[15] J. Redmon, S. Divvala, R. Girshick, and A. Farhadi, "You only look once: Unified, real-time object detection," in Proceedings of the IEEE conference on computer vision and pattern recognition, 2016, pp. 779-788.
[16] O. D. Trier and A. K. Jain, "Goal-directed evaluation of binarization methods," IEEE transactions on Pattern analysis and Machine Intelligence, vol. 17, no. 12, pp. 1191-1201, 1995.
[17] N. Otsu, "A threshold selection method from gray-level histograms," IEEE transactions on systems, man, and cybernetics, vol. 9, no. 1, pp. 62-66, 1979.
[18] G. Marcus, "Deep learning: A critical appraisal," arXiv preprint arXiv:1801.00631, 2018.
[19] S. J. Russell and P. Norvig, Artificial intelligence a modern approach. London, 2010.
[20] Y. LeCun, Y. Bengio, and G. Hinton, "Deep learning," nature, vol. 521, no. 7553, pp. 436-444, 2015.
[21] H. Lin, B. Li, X. Wang, Y. Shu, and S. Niu, "Automated defect inspection of LED chip using deep convolutional neural network," Journal of Intelligent Manufacturing, vol. 30, pp. 2525-2534, 2019.
[22] S. Saponara, A. Elhanashi, and Q. Zheng, "Developing a real-time social distancing detection system based on YOLOv4-tiny and bird-eye view for COVID-19," Journal of Real-Time Image Processing, vol. 19, no. 3, pp. 551-563, 2022.
[23] M. N. Kamel Boulos and E. M. Geraghty, "Geographical tracking and mapping of coronavirus disease COVID-19/severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) epidemic and associated events around the world: how 21st century GIS technologies are supporting the global fight against outbreaks and epidemics," vol. 19, ed: Springer, 2020, pp. 1-12.
[24] Z. Jiang, L. Zhao, S. Li, and Y. Jia, "Real-time object detection method based on improved YOLOv4-tiny," arXiv preprint arXiv:2011.04244, 2020.
[25] C.-Y. Wang, A. Bochkovskiy, and H.-Y. M. Liao, "YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors," in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023, pp. 7464-7475.
[26] J. Wang et al., "YOLO-Xray: A Bubble Defect Detection Algorithm for Chip X-ray Images Based on Improved YOLOv5," Electronics, vol. 12, no. 14, p. 3060, 2023.
[27] Z. Zou, K. Chen, Z. Shi, Y. Guo, and J. Ye, "Object detection in 20 years: A survey," Proceedings of the IEEE, 2023.
[28] L. Jiao et al., "A survey of deep learning-based object detection," IEEE access, vol. 7, pp. 128837-128868, 2019.
[29] J.-T. Lu et al., "Deep Spine: Automated lumbar vertebral segmentation, disc-level designation, and spinal stenosis grading using deep learning," in Machine Learning for Healthcare Conference, 2018: PMLR, pp. 403-419.
[30] A. Bochkovskiy, C.-Y. Wang, and H.-Y. M. Liao, "Yolov4: Optimal speed and accuracy of object detection," arXiv preprint arXiv:2004.10934, 2020.
[31] K. Simonyan and A. Zisserman, "Very deep convolutional networks for large-scale image recognition," arXiv preprint arXiv:1409.1556, 2014.
[32] K. He, X. Zhang, S. Ren, and J. Sun, "Deep residual learning for image recognition," in Proceedings of the IEEE conference on computer vision and pattern recognition, 2016, pp. 770-778.
[33] A. G. Howard et al., "Mobilenets: Efficient convolutional neural networks for mobile vision applications," arXiv preprint arXiv:1704.04861, 2017.
[34] M. Sandler, A. Howard, M. Zhu, A. Zhmoginov, and L.-C. Chen, "Mobilenetv2: Inverted residuals and linear bottlenecks," in Proceedings of the IEEE conference on computer vision and pattern recognition, 2018, pp. 4510-4520.
[35] W. Liu et al., "Ssd: Single shot multibox detector," in Computer Vision–ECCV 2016: 14th European Conference, Amsterdam, The Netherlands, October 11–14, 2016, Proceedings, Part I 14, 2016: Springer, pp. 21-37.
[36] K. He, X. Zhang, S. Ren, and J. Sun, "Spatial pyramid pooling in deep convolutional networks for visual recognition," IEEE transactions on pattern analysis and machine intelligence, vol. 37, no. 9, pp. 1904-1916, 2015.
[37] J. R. Uijlings, K. E. Van De Sande, T. Gevers, and A. W. Smeulders, "Selective search for object recognition," International journal of computer vision, vol. 104, pp. 154-171, 2013.
[38] U. Maulik and D. Chakraborty, "Remote Sensing Image Classification: A survey of support-vector-machine-based advanced techniques," IEEE Geoscience and Remote Sensing Magazine, vol. 5, no. 1, pp. 33-52, 2017.
[39] M. A. Hearst, S. T. Dumais, E. Osuna, J. Platt, and B. Scholkopf, "Support vector machines," IEEE Intelligent Systems and their applications, vol. 13, no. 4, pp. 18-28, 1998.
[40] J. Melton and A. R. Simon, Understanding the new SQL: a complete guide. Morgan Kaufmann, 1993.
[41] D. Arthur and S. Vassilvitskii, "K-means++ the advantages of careful seeding," in Proceedings of the eighteenth annual ACM-SIAM symposium on Discrete algorithms, 2007, pp. 1027-1035.
[42] C.-Y. Wang, I.-H. Yeh, and H.-Y. M. Liao, "YOLOv9: Learning What You Want to Learn Using Programmable Gradient Information," arXiv preprint arXiv:2402.13616, 2024.
[43] C.-T. Chien, R.-Y. Ju, K.-Y. Chou, and J.-S. Chiang, "YOLOv9 for Fracture Detection in Pediatric Wrist Trauma X-ray Images," arXiv preprint arXiv:2403.11249, 2024.
[44] A. Wang et al., "Yolov10: Real-time end-to-end object detection," arXiv preprint arXiv:2405.14458, 2024.
[45] J. Hosang, R. Benenson, and B. Schiele, "Learning non-maximum suppression," in Proceedings of the IEEE conference on computer vision and pattern recognition, 2017, pp. 4507-4515.