近年來,臺灣的農產業除了國內銷售,自國際銷售市場中獲利亦逐年上升。花卉產品為我國重要的外銷農產品之一,其中,蝴蝶蘭品種花卉的出口產值占總花卉產值逾六成的比例。國內業者在蝴蝶蘭盆苗出貨前,會透過其外觀型態進行品質篩選,過往中多以人力篩選,相當耗費時間以及人事成本。此外,隨著蝴蝶蘭銷售量成長、栽種數量增加,溫室規模擴大,在大型溫室內的微型氣候變化亦隨之加遽,環境差異皆可能影響蝴蝶蘭盆苗的生長品質。故如何快速地判別蝴蝶蘭盆苗生長狀態,以及掌握溫室內細微的環境因子變化,是相當重要的研究議題。 本研究導入物聯網(IoT)技術概念,於培植溫室內建置自動化蝴蝶蘭盆苗生長監控平台,此平台包含蝴蝶蘭盆苗生長性狀即時影像系統以及自動化溫室環境監控系統。前者即時影像系統主要透過機器視覺技術,建構出自動化影像機台,進行蝴蝶蘭盆苗的生長影像擷取,經過影像處理運算後,可針對盆苗之生長型態特徵,例如葉面積、葉幅等特徵進行估測,並將特徵資訊寫入後端資料庫內儲存生長參數。後者環境監測系統主要利用無線感測器網路(Wireless sensor networks, WSNs)技術建構出一套自動化環境監測網路,此網路具有可擴充性的動態路由,感測節點之間可藉由多跳式資料傳輸提升資料傳輸妥善率。此系統所監測之環境因子包含溫室內溫、濕度、陽光照度以及介質濕度等環境資訊。透過上述之二系統的整合,可自動化記錄蝴蝶蘭盆苗葉片的生長狀態,有助於減少人工量測盆苗生長狀態所耗費的時間,以及更準確監控溫室內生長環境的變化。 本研究將蝴蝶蘭盆苗即時影像系統,以及自動化環境監測系統架設於蘭園溫室內實驗測試,針對蝴蝶蘭紅花品種Dtps. I-Hsin Ice Coke “KH7359”共計420株3吋盆苗,分布於溫室內不同區域栽培,進行生長狀態辨識以及生長環境監測並進行分析,未來可供栽培業者生長環境的管理。
In recent years, Taiwan’s agricultural products target not only domestic sales but also international trades. Floriculture is one of the important agriculture products for exports. Phalaenopsis is the main floriculture sold in the international market. In order to ensure the quality of the flowers, farmers need to identify the best growth conditions of the Phalaenopsis seedlings by manually measuring the geometric characteristics. This task is time consuming and labor intensive. Furthermore, the scale of greenhouse production becomes larger when the sales go up, and this leads to the change in the atmospheric conditions and the negative impact on the quality of the seedlings. Therefore, it is very important to evaluate the quality of the seedlings by measuring their growth and monitor environmental factors through an efficient way and in a real-time manner. In this study, a remote real-time monitoring platform has been designed and implemented based on the IoT (Internet of Things) technology. Phalaenopsis seedlings were the research target. This platform included a real-time image system and an automated environmental monitoring system. For the former system, it used the vision technology to build an automatic machine. The growth status of Phalaenopsis seedlings was recorded through the images captured by the real-time image system. After image processing, the characteristics of Phalaenopsis seedlings, such as the leaf area and leaf width, were evaluated and the data were transferred to a back-end server. The latter system was implemented based on the wireless sensor networks technology. The environmental factors, such as temperature, humidity, lightness and medium moisture, were collected by sensor nodes. The sensor nodes transferred information through multiple paths and then built a dynamic routing network. With the integration of the two system, the time spent on manual measuring could be largely reduced and the greenhouse could be efficiently monitored. The proposed systems were used to monitor the Phalaenopsis (Dtps. I-Hsin Ice Coke “KH7359”). Phalaenopsis was cultivated in the same greenhouse under different environmental conditions. The historical monitoring data could be analyzed to find a better management strategy to greatly improve the quality of Phalaenopsis.