本研究致力於開發一套個體差異化自動澆灌系統，以解決現行蝴蝶蘭溫室澆灌過程中的澆灌點、澆水量不易控制之缺點。此自動澆灌系統架設於龍門型機器人上，透過植床影像規劃通過植株中心且垂直葉面方向的澆灌路徑，抵達澆灌點時，啟動噴嘴澆灌系統所設定的澆水量。實驗時龍門機器人上設有三組線性滑軌，以同時對多排植株進行澆灌，而每個線性滑軌皆配有一組流量比例閥組，其做為系統的澆灌設備，可藉由輸入電壓訊號精準地控制澆灌量。本研究的核心技術為多物件追蹤演算法，其藉由YOLOv4-tiny 物件偵測模型以及 SORT 演算法實現，系統應用此技術記錄各植株在追蹤期間的多幀均化姿態，並根據記錄的姿態規劃出澆灌路徑，線性滑軌會依澆灌路徑移動，以避開葉面區域進行澆灌，降低植株因葉面浸潤發生病害的機率。此外，本研究亦建立以 LoRa 通訊為基礎的介質溼度監控系統，監測植株介質的體積含水率，以做為澆灌量決策的依據。本研究自動澆灌系統實現了兩種澆灌模式，分別是精準澆灌模式和移動間斷水模式，精準澆灌模式表現良好的澆灌量控制能力，而移動間斷水模式則具有較高的澆灌效率，使用者可根據不同需求選擇澆灌模式。本研究以合作溫室提供的蝴蝶蘭植株為主要研究對象，然而本系統不受限於蝴蝶蘭植株上，未來仍可應用於其他澆灌目標，在實現精準澆灌的同時減少水資源的浪費

This research is devoted to developing an automatic and customized watering system for orchid seedlings to solve shortcomings of watering amount and location in the current greenhouse. This proposed system is installed on a gantry robot, and uses the image of the planting bed to plan the watering paths that pass through the center and perpendicular to the leaf direction. When the gantry robots move to the watering points, the nozzles will output the set amount of water. In experiments, the gantry robot is equipped with three linear slides to water multiple lines of plants simultaneously. Each linear slide is paired with a proportional valve as the watering equipment of the system, which can precisely control watering amount by the voltage signal according to the request of system. The core technology applied in this research is multiple object tracking, which is implemented by the object detection model YOLOv4-tiny and the SORT algorithms, which records the multi-frame average poses of all orchid seedlings, to be used to plan the watering paths. Finally, the linear slides will move along the paths to avoid the foliage area for watering, reducing the probability of plant diseases due to infiltration. In addition, a medium moisture monitoring system is established based on LoRa to observe the volumetric water content of the plant medium as a reference of the watering amount decision. Two watering modes are implemented on the watering system, namely precise watering mode and intermittent watering mode. The former mode is good in controlling the watering amount, while the later mode is focus on the watering efficiency. Users can choose the watering mode according to different needs. Although this research takes the orchid seedlings as the research object, the developed system can be applied to other plants to achieve precise watering, including the amount and location of water to reduce the waste of water resources and the grow plants better.

[1] 行政院農委會. (2020) 農業統計要覽. [Online]. Available: https:
//agrstat.coa.gov.tw/sdweb/public/book/Book.aspx, accessed: 2021-
12-09.
[2] 楊智凱、施瑩艷. (2016) 以智慧科技邁向臺灣農業 4.0 時代.
[Online]. Available: https://www.coa.gov.tw/ws.php?id=2505139,
accessed: 2021-12-09.
[3] 蔡宗顯. (2021) 財政統計通報 (第 22 號). [Online]. Available:
https://reurl.cc/dW01mD, accessed: 2021-12-09.
[4] R. Aronson. Farmbot, the open-source cnc farming robot. [Online].
Available: https://farm.bot/, accessed: 2022-05-25.
[5] J. Brown, K. Colombo, L. Salem, N. Jeng, R. Stothers, and S. Lees.
(July 11, 2017) Polar coordinate farm bot final project report. [Online].
Available: https://digitalcommons.calpoly.edu/imesp/205/, accessed:
2022-05-26.
[6] C. J. Choque Moscoso, E. M. Fiestas Sorogastúa, and R. S.
Prado Gardini, “Efficient implementation of a cartesian farmbot robot
for agricultural applications in the region la libertad-peru,” 2018 IEEE
ANDESCON, pp. 1–6, Cali, Colombia, Aug 22–24, 2018.
[7] M. Erick, S. Fiestas, R. Sixto, and G. Prado, “Modeling and
simulation of kinematics and trajectory planning of a farmbot cartesian
robot,” IEEE XXV International Conference on Electronics, Electrical
Engineering and Computing (INTERCON), pp. 1–4, Lima, Peru, Aug
08–10, 2018.
[8] T.-W. Chang, W.-C. Wang, and R. Chen, “Intelligent control system
to irrigate orchids based on visual recognition and 3d positioning,”
Applied Sciences, vol. 11, no. 10, 2021.
[9] T. S. Rajalakshmi, P. Panikulam, P. K. Sharad, and R. R. Nair,
“Development of a small scale cartesian coordinate farming robot with
deep learning based weed detection,” Journal of Physics: Conference
Series(JPCS), vol. 1969, no. 1, p. 012007, Jul 2021.
[10] L. Gao, M. Zhang, and G. Chen, “An intelligent irrigation system based
on wireless sensor network and fuzzy control,” Journal of Networks,
vol. 8, no. 5, pp. 1080–1087, May, 2013.
[11] Y.-C. Chang, T.-W. Huang, and N.-F. Huang, “A machine learning
based smart irrigation system with lora p2p networks,” 2019 20th AsiaPacific Network Operations and Management Symposium, pp. 1–4,
Matsue, Japan, Sep 18–20, 2019.
[12] I. K. Y. T. Permana, T. Mantoro, E. Irawan, D. O. Dwi Handayani,
and C. Safitri, “Increased efficiency of smart water systems for
vegetable plants using the deep learning classification approach,”
International Conference on Computing Engineering and Design, pp.
1–5, Singapore, Apr 11–13, 2019.
[13] T. N. Balasooriya, P. Mantri, and P. Suriyampola, “Iot-based smart
watering system towards improving the efficiency of agricultural
irrigation,” 2020 IEEE Global Conference on Artificial Intelligence
and Internet of Things (GCAIoT), pp. 1–7, Dubai, United Arab
Emirates, Dec 12–16, 2020.
[14] N. Ishak, A. H. Awang, N. Bahri, and A. Zaimi, “Gsm activated
watering system prototype,” 2015 IEEE International RF and
Microwave Conference (RFM), pp. 252–256, Kuching, Malaysia, Dec
14-16, 2015.
[15] B. N. Getu and H. A. Attia, “Automatic control of agricultural pumps
based on soil moisture sensing,” IEEE AFRICON, pp. 1–5, Addis
Ababa, Ethiopia, Sep 14–17, 2015.
[16] P. H. Tarange, R. G. Mevekari, and P. A. Shinde, “Web based automatic
irrigation system using wireless sensor network and embedded linux
board,” 2015 IEEE International Conference on Circuits, Power and
Computing Technologies [ICCPCT-2015], pp. 1–5, Nagercoil, India,
Mar 19–20, 2015.
[17] C. Kamienski, J.-P. Soininen, M. Taumberger, R. Dantas, A. Toscano,
T. Salmon Cinotti, R. Filev Maia, and A. Torre Neto, “Smart water
management platform: Iot-based precision irrigation for agriculture,”
Sensors, vol. 19, no. 2, p. 276, Jan 11, 2019.
[18] H. M. Yasin, S. R. Zeebaree, and I. M. Zebari, “Arduino based
automatic irrigation system: Monitoring and sms controlling,” 2019
4th IEEE Scientific International Conference Najaf, pp. 109–114, Al–
Najef, Iraq, Apr 29–30, 2019.
[19] Z. Muhammad, M. A. A. M. Hafez, N. A. M. Leh, Z. M. Yusoff,
and S. Abd Hamid, “Smart agriculture using internet of things with raspberry pi,” 2020 10th IEEE International Conference on Control
System, Computing and Engineering (ICCSCE), pp. 85–90, Penang,
Malaysia, Apr 21–22, 2020.
[20] X. Peng, Z. Mo, L. Xiao, and G. Liu, “A water-saving irrigation system
based on fuzzy control technology and wireless sensor network,”
2009 5th International Conference on Wireless Communications,
Networking and Mobile Computing, pp. 1–4, Beijing, China, Sep 24–
26, 2009.
[21] 李吉豪，” 溫室用龍門型澆水機器人研究”，國立清華大學動力
機械工程學系碩士論文， 2022 年 2 月。
[22] 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, Nov, 1998.
[23] R. Girshick, J. Donahue, T. Darrell, and J. Malik, “Rich feature
hierarchies for accurate object detection and semantic segmentation,”
Proceedings of the IEEE Conference on Computer Vision and Pattern
Recognition, pp. 580–587, Jun 23–28, 2014.
[24] R. Girshick, “Fast r-cnn,” 2015 IEEE International Conference on
Computer Vision (ICCV), pp. 1440–1448, Santiago, Chile, Dec 07–13,
2015.
[25] S. Ren, K. He, R. Girshick, and J. Sun, “Faster r-cnn: Towards real-time
object detection with region proposal networks,” IEEE Transactions on
Pattern Analysis and Machine Intelligence, vol. 39, no. 6, pp. 1137–
1149, Jun 01, 2017.
[26] J. Redmon, S. Divvala, R. Girshick, and A. Farhadi, “You only look
once: Unified, real-time object detection,” Proceedings of the IEEE
Conference on Computer Vision and Pattern Recognition, pp. 779–788,
Las Vegas, NV, USA, Jun 27–30, 2016.
[27] J. Redmon and A. Farhadi, “Yolo9000: better, faster, stronger,”
Proceedings of the IEEE Conference on Computer Vision and Pattern
Recognition, pp. 7263–7271, Honolulu, HI, USA, Jul 21–26, 2017.
[28] J. Redmon and A. Farhadi, “Yolov3: An incremental improvement,”
arXiv preprint arXiv:1804.02767, Apr 08, 2018.
[29] A. Bochkovskiy, C.-Y. Wang, and H.-Y. M. Liao, “Yolov4:
Optimal speed and accuracy of object detection,” arXiv preprint
arXiv:2004.10934, Apr 23, 2020.
[30] A. Bochkovskiy. (2020) Yolov4-tiny released: 40.2% ap50, 371
fps (gtx 1080 ti), 1770 fps tkdnn/tensorrt. [Online]. Available:
https://github.com/AlexeyAB/darknet/issues/6067, accessed: 2021-
10-22.
[31] W. Luo, J. Xing, A. Milan, X. Zhang, W. Liu, and T.-K. Kim, “Multiple
object tracking: A literature review,” Artificial Intelligence, vol. 293,
p. 103448, Apr, 2021.
[32] A. Bewley, Z. Ge, L. Ott, F. Ramos, and B. Upcroft, “Simple online
and realtime tracking,” 2016 IEEE International Conference on Image
Processing (ICIP), pp. 3464–3468, Phoenix, AZ, USA, Sep 25–28,
2016.
[33] N. Wojke, A. Bewley, and D. Paulus, “Simple online and realtime
tracking with a deep association metric,” 2017 IEEE International
Conference on Image Processing (ICIP), pp. 3645–3649, Beijing,
China, Sep 17–20, 2017.
[34] Semtech Corporation. What is LoRa. [Online]. Available: https:
//reurl.cc/GEE9dZ, accessed: 2022-05-08.
[35] LoRa Alliance. What is LoRaWAN. [Online]. Available: https:
//reurl.cc/5pp6oz, accessed: 2022-05-08.
[36] Stm32 nucleo f446re. [Online]. Available: https://reurl.cc/D33RvR,
accessed: 2022-05-08.
[37] J. Redmon. (2013–2016) Darknet:open source neural networks in c.
[Online]. Available: http://pjreddie.com/darknet/, accessed: 2021-10-
22.
[38] R. E. Kalman, “A New Approach to Linear Filtering and Prediction
Problems,” Journal of Basic Engineering, vol. 82, no. 1, pp. 35–45,
Jun 05, 1960.
[39] H. W. Kuhn and B. Yaw, “The hungarian method for the assignment
problem,” Naval Research Logistics, pp. 83–97, 1955.
[40] J. Hrisko. (2020) Capacitive soil moisture sensor theory, calibration,
and testing. [Online]. Available: https://reurl.cc/2mgDqa, accessed:
2022-05-08.
[41] K. Bernardin and R. Stiefelhagen, “Evaluating multiple object tracking
performance: the clear mot metrics,” EURASIP Journal on Image and
Video Processing, pp. 1–10, 2008.