自動載具應用於溫室葉菜株高自動化量測之系統整合

Systems Integration for Autonomous Vehicles Applications on Plant Height Measurement of Leafy Vegetables

劉柏誼(Po-I Liu)

指導教授 ： 顏炳郎

國立臺灣大學/生物資源暨農學院/生物機電工程學系/碩士(2022年)

https://doi.org/10.6342/NTU202200446

摘要

作物的株高在作物的生長模型扮演了一個重要的角色，株高數據可以幫助農民更好進行栽培與管理，或者避免植株因生長速度過快、氣候災害等原因而產生倒伏現象，又或者因應市場需求來管理收穫時作物的高度。不過作物高度量測往往需要大量的人力，因此本研究提出了一套結合導航、定位與自動量測株高的方法，並應用於溫室載具上。載具使用機器人作業系統進行系統整合，使用自適應蒙地卡羅定位法進行溫室中的定位，並且提出一套控制載具速度與角速度的方法，透過調整此方法中的控制參數，最後的導航誤差能夠在0.1 m以內。另外也設計了標定桿，搭配AR標誌，提供給相機辨識。相機辨識到AR標誌後會透過影像伺服控制的方式，將載具控制到標定桿的位置，影像伺服定位完成後，會進行株高的量測。株高量測的方法是透過深度相機，利用標定桿的高度、AR標誌的深度以及植物的深度等關係，來計算作物的高度。與人工量測株高的數據相比較，MAE為0.95 cm，RMSE為1.15 cm，MRE為16.60 %，R^2為0.80。

關鍵字

機器人作業系統 ； 載具定位與導航 ； 影像處理 ； 作物株高量測

並列摘要

The plant height of crops plays an important role in the growth model of crops. Plant height data can help farmers better cultivate and manage, or avoid plant lodging due to excessive growth rate, climate disasters, etc, or control the crop height at harvest to meet market demand. However, crop height measurement often requires a lot of manpower, so this study proposes a method that combines navigation, positioning and automatic plant height measurement, and applies it to the vehicle in greenhouse. The vehicle uses the Robotic Operating System (ROS) for system integration, uses the Adaptive Monte Carlo Localization (AMCL) method for positioning in the greenhouse, and proposes a set of methods to control the speed and angular velocity of the vehicle. By adjusting the control parameters in this method, the final navigation error can be within 0.1 m. In addition, a calibration rod is also designed, with an AR tag, which is provided for camera identification. After the camera recognizes the AR tag, it will control the vehicle to the position of the calibration rod through visual servo control. After the image servo positioning is completed, the plant height will be measured. The method of plant height measurement is to use the depth camera to calculate the height of the crop by using the height of the calibration rod, the depth of the AR tag and the depth of the plant. Compared with the data of manual measurement of plant height, MAE = 0.95 cm, RMSE = 1.15 cm, MRE = 16.60 %, and R^2 = 0.80.

並列關鍵字

Robotic Operating System ； vehicle positioning and navigation ； image processing ； crop height measurement

參考文獻

1. Bahman, L., Height Measurement of Basil Crops for Smart Irrigation Applications in Greenhouses using Commercial Sensors. 2019.

2. Blok, P.M., et al., Robot navigation in orchards with localization based on Particle filter and Kalman filter. Computers and Electronics in Agriculture, 2019. 157: p. 261-269.

3. Bohren, J. ROS wiki : smach. [cited 2022 02-03]; Available from: http://wiki.ros.org/smach.

4. Chang, A., et al., Crop height monitoring with digital imagery from Unmanned Aerial System (UAS). Computers and Electronics in Agriculture, 2017. 141: p. 232-237.

5. Chebrolu, N., T. Labe, and C. Stachniss, Robust Long-Term Registration of UAV Images of Crop Fields for Precision Agriculture. IEEE Robotics and Automation Letters, 2018. 3(4): p. 3097-3104.