蝴蝶蘭盆苗之幾何型態與螢光影像之分析

Abstract

目前臺灣為全世界前三大蝴蝶蘭出口國，並且其盆苗在市場上也佔有其一席之地，各國皆採用「接力栽培」的銷售模式，自臺灣進口成熟蝴蝶蘭種苗至國內栽培場所並進行催花，以節省自小苗種植至大苗所需之成本。但是，並非每一株所進口的盆苗皆可以催花成功，故該如何透過其形態特徵及其他快速的方式進行種苗的挑選，用以判別其種苗未來生長趨勢，是一項相當重要的課題。 目前有許多研究認為蝴蝶蘭其未來生長狀況可以透過其幾何型態進行評估，譬如其葉片面積、雙葉幅、葉片數或是莖徑等可透過觀察所得到的特徵，但是在一般的蘭園裡面種植著上百萬株的盆苗，若在出貨的時候逐一透過人力的方式進行型態特徵的估測，不但相當耗費人力及時間，並且在平常種植栽培的過程中，栽培人員也需要長期對種苗進行其型態特徵的估測，作為其栽培管理的依據，但是蘭園佔地廣大，管理人員需要在園內來回走動並對苗株進行檢驗，此方法不但耗費時間也會造成管理人員的疲累。 故本研究蝴蝶蘭盆苗為研究對象，並利用機器視覺技術建構自動化機台，並進行蝴蝶蘭盆苗的影像擷取，並透過影像處理進行其型態特徵，例如葉面積、葉幅及葉片數的估測。葉綠素螢光對環境具有高度靈敏性，目前有許多研究透過蝴蝶蘭螢光進行植物生長狀況的評估，並且螢光檢測具有快速、非破壞性檢測之優點，當發現植株生長情況未達預期時，栽培業者能隨時反應。利用機器視覺及機電整合等自動化技術建構本研究之影像監測平台，可減少人工量測時間的消耗，當監測資料被監測平台所量測到時，利用影像處理技術進行植株的生長參數量測，並結合數位資料庫管理系統進行生長參數的儲存，提供栽培業者即時資料庫供查閱，配合資料庫歷史資料之儲存，可提供栽培業者進一步做分析或生長環境的管理。

Taiwan now is the world’s top three exporters of Phalaenopsis, and the seedling pots of Phalaenopsis are also hot sales in this market. Most country follow a novel sales mode, relay cultivation, which imports mature seedling pots from Taiwan, and cultivate these pots into flowers in domestic cultivated field, with the advantage of costing down. However, not every seedling pot can be successfully induced to flower. Therefore, how to identify the growth of the seedling pots via measuring its geometric characteristic or other instant ways is a very important issue. Recently, there are many studies point out that Phalaenopsis’ growth status can be evaluated from the geometric characteristic of the pots such as leaf area, leaf width, the numbers of leaf or stem meter, etc. However, millions of seedling pots are cultivated in the field, it costs lots of time and human resource to evaluate the geometric characteristic of every pots when selling out these pots. On the other hand, farmers frequently evaluate the pots’ geometric characteristic as the information to management these pots. So, farmers should walk around and observe these pots in the wild field, and this way may causes farmers feel tired. In this study, we use Phalaenopsis seedling pots as our research object, and apply machine vision technology to build an automatic machine. After capturing seedling pot’s image, we further apply image processing to evaluate the geometric characteristic such as leaf area, leaf width and the numbers of leafs of the pot. Recently, many studies evaluate the plant’s growth via the chlorophyll fluorescence which has high sensitivity to the growth environment, and it can provide a fast and non-destructive measuring method. Farmers can react instantly as long as detect the growth of plants above standard. We develop a method based on machine vision and mechatronics technology, and it can reduce the manual measuring time. Once the monitoring data are measured, these data can be analysis via image processing method, and storage into the digital database which can provide farmers the history of the sensing data, furthermore, farmers can analysis these data to figure out a better management to cultivate the best quality of Phalaenopsis.