農業智慧系統引用於蝴蝶蘭抽梗管理之探討

面對臺灣蘭花產業應用人工智慧於蝴蝶蘭產業的瓶頸，本研究進行兩年試驗量測，探討蝴蝶蘭於不同溫室與不同年度運用人工智慧進行抽梗作業管理的可行情況，以累積低溫公式評估其運用於花梗長度之可行性，藉此調節蘭花上市銷售時間。累積低溫指標分成三型：荷蘭產業指標、台灣產業標準與試驗場主經驗標準。第一年試驗場域之量測試驗，發現試驗場域之蘭花花梗增長數據與累積低溫值之分佈關係，荷蘭與國內產業使用標準均為合理與可利用。而農場場主其夜溫10℃的使用標準也能說明花梗增長趨勢。第二年試驗場域受到大氣極端氣候影響，僅有農場場主夜溫10℃使用標準為合理。由於農業生產受到地域性與作物生物性等各種條件限制，智慧農業系統無法完全直接引入使用。結合作物生理基本原理與生產場域之相關經驗則可達到生產智慧管理之要求。

關鍵字

智慧農業；蝴蝶蘭；累積溫度；催梗

並列摘要

The Phalaenopsis industry in Taiwan's orchid industry faces the bottleneck of the application of artificial intelligence. This study conducted a two-years test and to explore the feasibility of using artificial intelligence to manage the stem lengthen of Phalaenopsis in different greenhouses and different years. The cumulative low temperature equation is used to evaluate the feasibility of applying it to predict the stem length, thereby the time for orchids to go on sale could be adjusted. There are three types of the cumulative low temperature inde: the Dutch industry index, the Taiwan industry standard and the owner's experience standard of the test site. The test field measurement experiment of the first year showed that the relationship between the growth data of orchid stem and the cumulative low temperature values was reasonable. The Dutch and domestic industry standards could be employed. The based night temperature of 10°C by the site owner can also be used to explain the growth trend of stems. In the second year, the test site was affected by the extreme climate, and only the farm owner's based night temperature of 10°C could be used as a reasonable standard. As agricultural production is restricted by various conditions such as regionality and crop biology, not all smart agricultural systems can be directly introduced and used. Integrating the basic principles of crop physiology and relevant experience in the production field can meet the requirements of intelligent production management.

並列關鍵字

Smart agriculture ； Phalaenopsis ； Accumulated temperature ； spiking

主題瀏覽