- 期刊
- OpenAccess
蝴蝶閆大苗選別系統之研制
Development of a sorting system for phalaenopsis seedlings
摘要
本研究主要為設計製造蝴蝶蘭大苗選別係統，由四部份組成：(1)取像定位機構、(2)影像辨識系統、(3)分級顯示裝置、(4)選別控制系統。此選別系統作業方式係將蝴蝶蘭大苗藉由置料平台與升降機構送至取像平台，利用影像辨識系統估算蝴蝶蘭大苗的幾何特徵及檢測病害，並根據選別標準進行分級，一次處理一株蝴蝶蘭大苗。選別流程由五個階段組成，第一階段為進料：將蝴蝶蘭大苗放置在取像定位機構之置料平台，啓動升降機構，降低置料平台，將大苗送至取像平台。第二階段為擷取上視影像：驅動頂部CCD攝影機擷取蝴蝶蘭大苗上視影像。第三階段為擷取前視影像：驅動伺服馬達轉動取像平台，驅動前方CCD攝影機擷取蝴蝶蘭大苗前視影像。第四階段為選別：檢測大苗是否罹患病害，並估算蝴蝶蘭大苗幾何特徵值，根據蝴蝶蘭大苗分級標準進行選別，顯示選別結果。第五階段為退料：啓動升降機構，升高置料平台，將蝴蝶蘭大苗取出。 根據台糖公司外銷的選別標準，本研究使用此選別系統與人工方式，針對430株蝴蝶蘭大苗分別進行機器與人工選別作業，由試驗結果顯示，機器選別的正確率為90%，每一株蝴蝶蘭大苗的平均選別作業時間為21.15 sec；人工選別的正確率則為97.2%，每一株蝴蝶蘭大苗的平均選別作業時間為27.42 sec；機器選別的速度較人工選別約快22.3%。
並列摘要
The objective of this study is to design a sorting system for Phalaenopsis seedlings. The sorting system consists of four major parts:(1)image grabbing and positioning mechanism,(2)pattern recognition system,(3)display panel, and (4) control system. The system handles one pot of plant at a time. Five stages of the sorting process are summarized as follows. Stage for loading: put one plant pot on the table of the image grabbing and positioning mechanism, and activate the lifting mechanism to lower the table. Stage for taking top image: trigger CCD camera on the top to take image from the top view of the plant. Stage for taking front image: activate a servomotor to rotate the circular plate of image grabbing and positioning mechanism (the rotation angle depends on the top image), and trigger CCD camera on the front to take image from the front view of the plant. Stage for sorting: detect the seedling disease, analyze and estimate the values of geometric characteristic, sort the Phalaenopsis plant according to the given sorting standard, and display the result. Stage for unloading: activate the lifting mechanism to lift the table, and unload plantpot from the table. We used 430 pots of plants of Phalaenopsis seedlings (Dtps. Taisuco Gallantry) to test the sorting system. The results were compared to manual sorting. According to the results, we were able to achieve a sorting rate of 21.15 sec/pot compared to27.42 sec/pot of manual sorting, to accuracy of 90% compared to 97.2%. Our machine can save up to 22.3% time of totally handling by man.