本研究成功開發出一個連續式尿液檢測試劑判讀系統，此系統使用輸送帶進行試劑的運送，並利用USB Camera擷取影像，將影像經由程式來判讀尿液十項檢測試劑的檢測結果。
為了提高尿液檢測試劑的精確度並降低判斷誤差，本研究開發了一個影像辨識系統，能夠自動分析尿液檢測試劑的顏色變化，此系統使用Python 程式撰寫，進行影像預處理，如：提高對比、輪廓辨識，並利用色彩校正、HSV 色彩空間顏色比對概念，進行試劑的判讀以及檢測，並透過多執行緒處理以及引入輸送帶達到快速且大量的判讀。
為了確認本系統的光源環境以及輸送帶速率等系統參數達到最佳化效果，本研究建立了可模擬特定光強度的暗箱環境進行實驗，經實驗確認系統參數。模擬試紙以及真實檢體試劑的實驗結果顯示，在55格色塊中，有1格色塊正確率為88.3%，6格色塊正確率介於95%~98.3%之間，其餘48 格色塊之正確率皆為100%，數據可顯示本研究所開發之連續式尿液檢測試劑判讀系統可正確還原色彩並具有一定的精確度。

This study successfully developed a continuous urine test strip interpretation system. The system utilizes a conveyor belt for transporting test strips and is equipped with a touchscreen for user interaction. Users can simply place the test strips on the conveyor belt, and the system can automatically recognize,analyze, and display the test results through the interface.
To enhance the accuracy of urine test strip interpretation and reduce judgment errors, the study implemented an image recognition system.This system,written in Python,employs concepts such as image preprocessing, contrast enhancement, contour recognition, color correction, HSV color space and color matching. It achieves rapid and high-volume interpretation through multithreaded processes and the integration of a conveyor belt.
To ensure optimal system performance regarding light source conditions and conveyor belt speed,the study conducted experiments in a simulated dark box environment.The experimental results,using both simulated and real specimen test strips, demonstrate that the developed automated urine test strip interpretation system accurately reproduces colors and exhibits a certain level of accuracy.

[1] 周士閔（2024年3月22日）。尿尿完看一下！10種尿液顏色狀況看健康。泌尿健康。https://helloyishi.com.tw/urological-health/health-concerns-related-to-urine-color/
[2] 盧映慈（2020年1月8日）。抽血正常腎功能卻剩一半！醫師：腎臟出問題從「尿」開始。Heho健康。https://heho.com.tw/archives/63259
[3] 李赫（2019年4月11日）。尿為什麼是黃色的？尿液到底能不能喝？。PanSci 泛科學。https://pansci.asia/archives/156869
[4] 科學Online（2010年8月24日）。水分攝取與尿液形成的調節 (Urine Formation)-上。https://highscope.ch.ntu.edu.tw/wordpress/?p=6884
[5] Chien, T. I., Kao, J. T., Liu, H. L., Lin, P. C., Hong, J. S., Hsieh, H. P., & Chien, M. J. (2007). Urine sediment examination: a comparison of automated urinalysis systems and manual microscopy. Clinica Chimica Acta, 384(1-2), 28-34. https://doi.org/10.1016/j.cca.2007.05.012
[6] 陳韻安（2012）。龍眼花，殼，種子，葉與枝椏萃取物之降尿酸研究〔未出版之碩士論文〕，國立臺灣大學獸醫學研究所。
[7] 姬軍，王丹，王云龍，蔣昌松，袁青（2015）。CIELAB 色彩空間內視鏡色彩還原能力比較分析。中國測試，41（5）。
[8] Rodriguez, M. (1995). “A graphic arts perspective on RGB-to-CMYK conversion,” International Conference on Image Processing, 2, pp. 319-322. https://doi.org/10.1109/icip.1995.537479
[9] Chang, Y. C., & Reid, J. F. (1996). RGB calibration for color image analysis in machine vision. IEEE Transactions on image processing, 5(10), 1414-1422. https://doi.org/10.1109/83.536890
[10] IEEM Wiki（n.d.）。Minecraft 像素畫色彩使用指南/基礎知識。https://reurl.cc/09j5Ob
[11] 中國LED網（2015年8月13日）。道路照明標準將進入中間視覺的評價時代。https://kknews.cc/zh-tw/news/lp5vlag.html
[12] Proykova, A., Samaras, T., Ion, R-M., Bruzell, E., Doré, J-F., Nicolo, M., O’Hagan, J., Sánchez-Ramos, C., Kerkhof, L. (2018). Potential Risks Human Health LEDs.
[13] 王仁宏（2019）。利用時間特徵來辨別由聽視覺一致與聽視覺不一致之刺激所引發之功能性核磁共振影像信號〔未出版之碩士論文〕，國立臺灣大學生醫電子與資訊學研究所。
[14] Longuet-Higgins, H. C. (1981). A computer algorithm for reconstructing a scene from two projections. Nature, 293(5828), 133-135. https://doi.org/10.1038/293133a0
[15] 衛生福利部統計處（2023年6月12日）。111年國人死因統計結果。https://www.mohw.gov.tw/cp-16-74869-1.html
[16] ARDUINO UNO R3 (n.d.). UNO R3. https://docs.arduino.cc/hardware/uno-rev3
[17] Raspberry Pi (n.d.). Raspberry Pi 4 Tech Specs. https://www.raspberrypi.com/products/raspberry-pi-4-model-b/specifications/
[18] BeagleBone® Black (n.d.). What is BeagleBone® Black?. https://www.beagleboard.org/boards/beaglebone-black
[19] Yu, J. (2023). Based on Gaussian filter to improve the effect of the images in Gaussian noise and pepper noise. Journal of Physics: Conference Series, 2580(1), 012062. https://doi.org/10.1088/1742-6596/2580/1/012062
[20] Canny, J. (1987). A computational approach to edge detection. Readings in Computer Vision, 184-203.
[21] Xu, Z., Baojie, X., & Guoxin, W. (2017). Canny edge detection based on Open CV. In 2017 13th IEEE international conference on electronic measurement & instruments (ICEMI). https://doi.org/10.1109/icemi.2017.8265710
[22] Manuaba, I. B. P., Indah, K. A. T., Fahmi, M., & Salsabila, I. N. (2022). An improvement object detection method findcontour with fuzzy logic for detect balinese script object. Aptisi Transactions on Technopreneurship (ATT), 4(3), 257-262. https://doi.org/10.34306/att.v4i3.264
[23] Mishra, S., & Thanh, L. T. (2022). SATMeas - Object Detection and Measurement: Canny Edge Detection Algorithm. In International Conference on AI and Mobile Services (pp. 91-101). Cham: Springer International Publishing.
[24] Mangin, J.-F. (n.d.). Entropy minimization for automatic correction of intensity nonuniformity. Proceedings IEEE Workshop on Mathematical Methods in Biomedical Image Analysis. MMBIA-2000. https://doi.org/10.1109/mmbia.2000.852374
[25] 趙曉晴，劉景鑫，張海濤，胡笑含，李大軍，李慧盈（2019）。色彩空間變換和基於距離變換的分水嶺演算法在白血球影像分割中的應用。中國醫療設備，34（7），5-9。
[26] Dey, N., Ashour, A., Shi, F., & Balas, V. E. (2018). Soft computing based medical image analysis. San Diego: Elsevier Science.
[27] Malakhov, A. (2016). Composable multi-threading for python libraries. Proceedings of the Python in Science Conference.
[28] Tahir Rasheed, M., & Shi, D. (2023). Towards low-light image restoration via color correction matrix learning. Computer Science and Technology. https://doi.org/10.57237/j.cst.2023.03.003