透過您的圖書館登入
IP:54.152.77.92
  • 學位論文

多源數據結合作物模式應用於蘆筍智慧生產

Integration of Multiple Data Resources and Crop Model for Asparagus Smart Production

指導教授 : 劉力瑜

摘要


蘆筍主要採收部位為其嫩莖,臺灣以留母莖法栽培蘆筍、且近年設施栽培比例提升,與蘆筍原生之溫帶地區的種植方法不同。同時,近年政府力推智慧農業政策,期望透過物聯網(Internet of Things, IoT)感測系統、數據科技建立之作物模式(Crop Model)等技術整合,協助農業提高生產效率及產值。然而,在科技導入的過程中可能遇到多種問題導致資料缺失而無法用以建模,因此本研究考量環境監測資料的來源種類,開發可適用於臺灣蘆筍栽培方式的作物模式,目標預測嫩莖高度生長,用以規劃智慧生產排程並以利後續的產量預估。 本研究利用作物模式方法建立嫩莖高度生長模式,並比較多種非線性生長模式(Nonlinear Growth Model)以及納入不同自變數的差異,最終依據不同的資源需求打造客製化生長模擬預測。結果發現,將自變數設為「積溫」的Gompertz生長模式對於適配嫩莖的生長有良好表現;其中,若透過開放的室外測站資料以及多元線性迴歸模型模擬溫室內氣溫,其可用於計算出模型自變數的積溫,而且所建立之生長模式的擬合度與直接以實際氣溫建立的模式接近。 另外,研究亦發現所開發之生長模式可適用於不同季節、不同場域中,嫩莖生長速率在各種情況下皆無顯著差異。雖此模式在實際應用中仍無法精準收斂嫩莖的預期採收時間,但已可更清楚地掌握不同季節與氣溫下採收時間的差異,並提出可能的採收時間區間。因此,依據本研究的作物模式,可提供農戶依據不同IoT資源及數據來源選擇最具落地可行性的嫩莖高度生長模式,用於協助蘆筍產業走向智慧化生產。

並列摘要


Taiwan has cultivated asparagus with the mother stalk cultural method, which is different from the cultivation method in temperate regions. The facility cultivation system for asparagus cultivation has increased in recent years. Meanwhile, the government has promoted smart agriculture policies to help farmers improve production efficiency and value through the integration of technologies such as the Internet of Things (IoT) system and the Crop Model established by data-driven technology. Difficulties may occur when integrating modern technologies, for example, the missing data caused by the deficiency of IoT systems. This research considered the sources of environmental monitoring data and developed crop models that can be applied to Taiwan's asparagus cultivation methods. The goal was to predict the growth of spear which is the main harvest part of asparagus, in order to schedule production in advance and facilitate subsequent yield estimation. The research estimated spear height via different nonlinear growth models by adapting various explanatory variables. We also predicted height based on different environmental factors being enquired from different resources. The results showed the Gompertz growth model with "thermal time" as the explanatory variable fits well to spear growth. If the "thermal time" variable was replaced as the simulated greenhouse temperature from the outdoor weather station with the multiple linear regression model, it could also predict the spear growth well. The study found that the growth model can be applied to different seasons and different fields. In conclusion, we hoped our study can help farmers to choose the most feasible spear height growth model based on different data resources and the asparagus industry move towards smart production.

參考文獻


Archontoulis, S. V., Miguez, F. E. (2015). Nonlinear Regression Models and Applications in Agricultural Research [Article]. Agronomy Journal, 107(2), 786-798. https://doi.org/10.2134/agronj2012.0506
Armstrong, J. S., Collopy, F. (1992). Error measures for generalizing about forecasting methods: Empirical comparisons. International journal of forecasting, 8(1), 69-80.
González Sánchez, A., Frausto Solís, J., Ojeda Bustamante, W. (2014). Predictive ability of machine learning methods for massive crop yield prediction.
Graefe, J., Heissner, A., Feller, C., Paschold, P.-J., Fink, M., Schreiner, M. (2010). A process-oriented and stochastic simulation model for asparagus spear growth and yield. European journal of agronomy, 32(3), 195-204.
Heißner, A., Schmidt, S., Schonhof, I., Feller, C., Schreiner, M. (2006). Spear yield and quality of white asparagus as affected by soil temperature. European journal of agronomy, 25(4), 336-344.

延伸閱讀