為了解決當前台灣從事農業人口短缺之問題，本研究開發一套蘭花開花株品質分級自動化辨識系統。首先建立葉片損傷、花朵顏色及花苞、花朵各種損傷之數千張影像資料庫，藉由深度學習中不同的物件偵測模型訓練，辨識蘭花開花株之不同等級並分類。在利用影像資料庫訓練完後，此辨識系統可以藉由攝影機取像，並比對瑕疵之特徵，即時分辨蘭花開花株是否於葉片及花朵上有瑕疵，進一步對於切花及盆花的花朵和花苞進行品質分級。本研究所使用之深度學習影像訓練集皆於合作蘭花種植園區實地拍攝的影像，藉由所開發之自動化分流機構，整合影像辨識成果，可達成蘭花開花株品質分級自動化辨識之目的，符合產業上的需求。
本研究開發一套圖形使用者介面，將訓練完畢的深度學習物間偵測演算法與PyQt5進行結合，設計可以即時監控影像辨識之結果，並將重要的資訊顯示在訊息欄，此介面可以讓使用者方便觀看影像辨識成果，對於產業實際應用上有更好的使用效果。

In order to solve the shortage of agricultural workers in Taiwan nowadays, this research aims to develop an automatic visual recognition system for the quality classification of orchid, planted in a greenhouse. In the system, thousands of image databases, including leaf damage, flower color & bud, and flower defect, have been established to identify the defects of orchids and to classify different grades of orchid by training different deep learning object detection algorithms. After training the databases, use the built camera in this recognition system to take pictures, which then be used to compare the features of image to instantly identify the orchid defect on the leaves and flowers, or to classify the quality classification of cut flowers and potted by flowers and buds. In this research all pictures used for the training of deep learning were photographed in the cooperative orchid garden. As a result, the system is reliable when is used in any orchid plant company. To meet the needs of the industry, a simple prototype of an automated shunting mechanism has been made to prove the feasible of the automatic identification of the quality classification for orchid. Integrating the results of image recognition, it can deliver the single item of orchid to the assigned destination.
Combining the trained deep learning object detection algorithms with PyQt5, a graphical user interface is designed for the friendly use of real-time monitoring of image recognition and for displaying the important information in the message bar.

[1] https://www.thenewslens.com/article/126192, February, 2020.
[2] https://statview.coa.gov.tw/aqsys_on/importantArgiGoal_lv3_1_ 6_3_1.html, February, 2020.
[3] https://agrstat.coa.gov.tw/sdweb/public/trade/tradereport.aspx, February, 2020.
[4] Y. LeCun, L. Bottou, Y. Bengio, and P. Haffner, "Gradient-based learning applied to document recognition," Proceedings of the IEEE, vol. 86, no. 11, pp. 2278-2324, 1998.
[5] A. Krizhevsky, I. Sutskever, and G. E. Hinton, "Imagenet classification with deep convolutional neural networks," Neural Information Processing Systems Conference, pp. 1097-1105, Les Vegas, USA, December 3-8, 2012.
[6] K. Simonyan and A. Zisserman, "Very deep convolutional networks for large-scale image recognition," Proc. Int. Conf. Learn. Represent., pp. 1-14, 2015.
[7] K. He, X. Zhang, S. Ren, and J. Sun, "Deep residual learning for image recognition," The IEEE Conference on Computer Vision and Pattern Recognition, pp. 770-778, Las Vegas, USA, June 26 - July 1, 2016.
[8] R. K. Srivastava, K. Greff, and J. J. a. p. a. Schmidhuber, "Highway networks," Machine Learning, May 3, 2015.(https://arvix.org/1505.00387.pdf)
[9] R. Girshick, J. Donahue, T. Darrell, and J. Malik, "Rich feature hierarchies for accurate object detection and semantic segmentation," The IEEE Conference on Computer Vision and Pattern Recognition, pp. 580-587, Columbus, USA, June 24-27, 2014.
[10] R. Girshick, "Fast r-cnn," The IEEE International Conference on Computer Vision, pp. 1440-1448, Santiago, Chile, 2015.
[11] S. Ren, K. He, R. Girshick, and J. Sun, "Faster r-cnn: Towards real-time object detection with region proposal networks," Neural Information Processing Systems Conference, pp. 91-99, Montreal, Canada, December 7-12, 2015.
[12] J. Redmon, S. Divvala, R. Girshick, and A. Farhadi, "You only look once: Unified, real-time object detection," The IEEE Conference on Computer Vision and Pattern Recognition, pp. 779-788, Les Vegas, USA, June 26 - July 1, 2016.
[13] J. Redmon and A. Farhadi, "YOLO9000: better, faster, stronger," The IEEE Conference on Computer Vision and Pattern Recognition, pp. 7263-7271 , Honolulu, Hawaii, July 22-25,2017.
[14] J. Redmon and A. Farhadi, “YOLOv3: An Incremental Improvement,” CoRR, April, 2018. (https://arxiv.org/pdf/1804.02767.pdf)
[15] Z. Cai and N. Vasconcelos, "Cascade r-cnn: Delving into high quality object detection," The IEEE Conference on Computer Vision and Pattern Recognition, pp. 6154-6162, Utah, USA, June 19-21, 2018.
[16] M. T. bin MohamadAzmi and N. M. Isa, "Orchid disease detection using image processing and fuzzy logic," International Conference on Electrical, Electronics and System Engineering, pp. 37-42, Kuala Lumpur, Malaysia, December 4-5, 2013.
[17] W. M. N. Wan Mohd Fadzil, M. S. B. S. Rizam, R. Jailani, and M. T. Nooritawati, "Orchid leaf disease detection using border segmentation techniques," IEEE Conference on Systems, Process and Control (ICSPC 2014), pp. 168-173, Kuala Lumpur, Malaysia, December 12-14, 2014.
[18] A. Fuentes, S. Yoon, S. C. Kim, and D. S. Park, "A robust deep-learning-based detector for real-time tomato plant diseases and pests recognition," Sensors, vol. 17, no. 9, p. 2022, 2017.
[19] 林采蓉， "基於深度學習之自動化蘭花苗株病變視覺辨識系統"，國立清華大學動力機械工程學系碩士論文，民國108年7月。
[20] https://www.logitech.com/zh-tw/product/c922-pro-stream-webcam?crid=34, February, 2020.
[21] https://medium.com/@yehjames/%E8%B3%87%E6%96%99%E 5%88%86%E6%9E%90-%E6%A9%9F%E5%99%A8%E5%AD%B8%E7%BF%92-%E7%AC%AC%E4%B8%80%E8%AC%9B-python%E6%87%B6%E4%BA%BA%E5%8C%85-anaconda-%E4%BB%8B%E7%B4%B9-%E5%AE%89%E8%A3%9D-f8199fd4be8c, February, 2020.
[22] G. Bradski and A. Kaehler, Learning OpenCV, O'Reilly Media, Inc., 2008.
[23] https://pypi.org/project/PyQt5/, February, 2020.
[24] https://blog.csdn.net/rosefun96/article/details/79442814, February, 2020.
[25] https://en.wikipedia.org/wiki/AlphaGo, February, 2020.
[26] https://www.gigabyte.com/tw/Graphics-Card/GV-N208TTURBO-11GC-rev-10#kf, February, 2020.
[27] https://www.leadtek.com/cht/products/AI_HPC(37)/NVIDIA_Jets on_TX2(10782)/detail, February, 2020.
[28] http://tensorflowkeras.blogspot.com/2017/08/keras.html, February, 2020.
[29] https://www.itread01.com/content/1542587896.html, February, 2020.
[30] https://www.tensorflow.org/guide/graphs, February, 2020.
[31] https://blog.csdn.net/bestrivern/article/details/89433023, February, 2020.
[32] https://github.com/tzutalin/labelImg, February, 2020.
[33] https://store.arduino.cc/usa/mega-2560-r3, February, 2020.
[34] https://goods.ruten.com.tw/item/show?21911302950070#info, July, 2019.
[35] https://www.hcmat.tw.bysources.com/?fbclid=IwAR1VXh_sWFR OoLvAqOchGUvR8wk9VXUVzD6vIzX5hq4YGVUNg4yhb5hydHw, December, 2018.
[36] https://global.sharp/products/device/lineup/data/pdf/datasheet/g p2y0a21yk_e.pdf, February, 2020.
[37] 皇基股份有限公司內部資料，2020。
[38] L. Y. Pratt, "Discriminability-based transfer between neural networks," Neural Information Processing Systems Conference, pp. 204-211, Denver, USA, 1993.
[39] https://chtseng.wordpress.com/2017/11/11/data-augmentation-%E8%B3%87%E6%96%99%E5%A2%9E%E5%BC%B7/, July, 2018.
[40] https://medium.com/@chih.sheng.huang821/%E6%B7%B1%E5 %BA%A6%E5%AD%B8%E7%BF%92%E7%B3%BB%E5%88%97-%E4%BB%80%E9%BA%BC%E6%98%AFap-map-aaf089920848, February, 2020.
[41] https://github.com/open-mmlab/mmdetection/blob/master/docs/MODEL_ZOO.md, February, 2020.