智慧農業機器人可藉由無人載具搭配農業自動化設備來大幅降低農業勞動成本，幫助緩解農業勞動人口的缺口與提升整體蔬果的經濟產量。本論文主要是基於無人地面載具(unmanned ground vehicles，UGV)來開發具精準3D定位之智慧雷射系統，以達到快速物理除蟲目的，進而促進智慧農業的發展。此系統分為三大區塊：嵌入式深度學習辨識害蟲影像模組、藉由深度相機之精準3D定位系統、雷射光束除蟲掃描系統。目前可達到在雷射工作距離40~70 cm的下，深度(z軸)誤差小於0.5%，xy平面誤差小於2 mm。另因採用的結構光相機易受太陽光的影響，可藉由自動調整曝光參數，強化室外可用性。而雷射光束能夠快速精準地照射在害蟲頭部位置上，在距離50 cm時可控制系統誤差小於2 mm。考量搭載於UGV下之電力有限，經實驗證明雷射照射在害蟲頭部致其喪失吃葉子功能所需雷射劑量最低。智慧影像辨識及定位於NVidia TX2系統下可達到20幀率(frame rate)，且24 W波長為445 nm之雷射經振鏡將雷射光束對準於害蟲頭部並曝光150 ms，可有效在不須立刻殺死害蟲情況下，使害蟲不再進食，並於兩天內死亡，達到節省電力下高效率物理控制害蟲數量。

In this thesis, an artificial intelligent laser pest system with precise 3D targeting on the UGV has been developed to achieve the goal of rapid physical pest removal, and then promote the development of intelligent agriculture. The system includes three parts:embedded system module for deep learning with pest, precise 3D positioning system by depth camera, and laser pest controller system. Currently, the depth (i.e. z axis) error is less than 0.5% and the xy plane error is less than 2 mm at the working distance about 40-70 cm.
Because the structure light camera is easy to be interfered by the sun light, the outdoor availability is enhanced by auto adjusting the exposure parameters. The laser can target to the pest’s head quickly and accurately, and the positioning error of the system is controlled less than 2 mm when the distance is 50 cm. Furthermore, due to the limit of the power source in UGV, the laser targeting on the pest’s head is the most effective with minimum laser dosage. Intelligent image recognition and positioning in NVIDIA TX2 system can achieve the frame rate of 20 Hz, and the laser with 24 W and 445 nm can be targeted at the head of the pest and exposed for 150 ms by Galvo scanners. It can effectively prevent the pest from eating and die within two days.

[1] https://www.ly.gov.tw/Pages/Detail.aspx?nodeid=6590&pid=85487
[2] https://www.coa.gov.tw/ws.php?id=2506373
[3] K. Kurashiki, T. Fukao, K. Ishiyama, T. Kamiya, and N. Murakami, “Orchard Traveling UGV Using Particle Filter Based Localization and Inverse Optimal Control,” IEEE/SICE International Symposium on System Integration (2010) 21.
[4] R. Tazzari, D. Mengoli, and L. Marconi, “Design Concept and Modelling of a Tracked UGV for Orchard Precision Agriculture,” IEEE International Workshop on Metrology for Agriculture and Forestry (2020) 4.
[5] H. S. Abdullahi, Ray E. Sheriff, and F. Mahieddine, “Convolution Neural Network in Precision Agriculture for Plant Image Recognition and Classification,” IEEE Seventh International Conference on Innovative Computing Technology (2017)16.
[6] C. Song, C. Wang, and Y. Yang, “Automatic Detection and Image Recognition of Precision Agriculture for Citrus Diseases,” IEEE Eurasia Conference on IOT, Communication and Engineering (2020) 23.
[7] https://www.coa.gov.tw/theme_list.php?theme=photo_news&class=1&id=934
[8] https://www.newsmarket.com.tw/herbcides/ch02/
[9] https://www.newsmarket.com.tw/blog/111984/
[10] 曹智鈞，GPS 應用於桿式噴藥車變率施噴系統之實測，碩士論文，台灣大學生物產業機電工程學系(2004)。
[11] 謝環宇，適用於果實採摘之六軸繩索機械臂機構設計與運動模擬分析，碩士論文，台灣大學生物產業機電工程學系，2014。
[12] I. Kandori, K. Tsuchihara, T. A. Suzuki, T. Yokoi, and D. R. Papaj, “Long Frontal Projections Help Battus Philenor (Lepidoptera: Papilionidae) Larvae Find Host Plants,” PLoS ONE 10 (2015) e0131596.
[13] https://www.nvidia.com/zh-tw/autonomous-machines/embedded-systems/jetson-tx2/
[14] M. Quigley, K. Conley, B. P. Gerkey, J. Faust, T. Foote, J. Leibs, R. Wheeler, and A. Y. Ng, “ROS: An Open-Source Robot Operating System,” Workshops at the IEEE International Conference on Robotics and Automation (2009).
[15] W. Liu, D. Anguelov, D. Erhan, C. Szegedy, S. Reed, C. Y. Fu, and A. C. Berg, “SSD: Single Shot Multibox Detector,” European Conference on Computer Vision (2016) 21.
[16] J. Redmon, S. Divvala, R. Girshick, and A. Farhadi, “You Only Look Once: Unified, Real-Time Object Detection,” IEEE Conference on Computer Vision and Pattern Recognition (2016) 214.
[17] J. Redmon and A. Farhadi, “YOLOv3: An Incremental Improvement,” arXiv:1804.02767 (2018).
[18] C. Chen, M. Y. Liu, O. Tuzel, and J. Xiao, “R-CNN for Small Object Detection,” Asian Conference on Computer Vision (2016).
[19] W. Chen, Y.-Q. Shi, and G. Xuan, “Identifying Computer Graphics using HSV Color Model and Statistical Moments of Characteristic Functions,” IEEE International Conference on Multimedia and Expo (2007) 5.
[20] R. C. Gonzalez and R. E. Woods, Digital Image Processing, 4th Edition (2018).
[21] A. David and S. Vassilvitskii, “K-means++: The Advantages of Careful Seeding,” SODA ‘07: Proceedings of the Eighteenth Annual ACM-SIAM Symposium on Discrete Algorithms (2007) 1027.
[22] Q. Ji, and Y. Zhang, “Camera Calibration with Genetic Algorithms,” IEEE Transactions on Systems, Man, and Cybernetics-Part A: Systems and Humans 30 (2001) 120.
[23] M. A. Sutton, J. J. Orteu, and H. W. Schreier, Image Correlation for Shape, Motion and Deformation Measurements: Basic Concepts, Theory and Applications, Springer Verlag (2009).
[24] H. W. Yoo, S. Ito, and G. Schitter, “High Speed Laser Scanning Microscopy by Iterative Learning Control of a Galvanometer Scanner,” Control Eng. Pract. 50 (2016) 12.
[25] A. Gray, A Treatise on Gyrostatics and Rotational Motion, MacMillan, London (1918).