使用Objet RP快速原型機時，加工中曾多次出現工件建構形狀錯誤或缺陷；但系統仍未察覺而持續加工，導致造成工件外型損毀，浪費材料和時間。本研究主要針對此問題，發展一套屬於Objet RP快速原型機的線上影像監控檢測系統，以便在加工中，即時辨別工件外型輪廓是否有缺陷產生，並能及時發出警報。
本系統包含視覺取像模組、PLC程式控制模組、機構組立模組及影像檢測軟體模組四個部分。視覺取像模組由CCD攝影機、鏡頭、影像擷取卡及照明系統所構成，用以擷取影像。PLC程式控制模組中，光感測器藉由噴頭移動時的觸動，下達訊號至PLC，使PLC傳送擷取影像的訊號，此外當RP系統發生故障時，PLC亦可控制警示器。機構組立模組用以支撐視覺取像模組，以利取像模組做固定位置的取像作業。影像檢測軟體模組針對Objet RP採用兩種影像辨識的方法，分別為面積比較法及量測點座標間距法；此兩種方法一起使用，可以有效辨別幾何外型輪廓及其缺陷。所使用的面積比較法之比例值設定為0.95至1.05之間，量測座標間距法之最大容許間距設定為10 pixels。運用在本論文中所設計的模型，能夠有效的區別外型輪廓的缺陷有無，並能在發現缺陷時，即時發出警訊。

Based on the experiences of using Objet Rapid Prototyping (RP) machine, some defects would happen during the fabrication. However, the Objet RP system could not detect the existence of the defects and allowed the process to continue, resulting in defective models and waste of the materials and time. To solve this problem, this research developed an on-line visual monitoring system for Object RP machine which can immediately identify the defects of the shape in each layer during the process and can initiate the alarm system if a defect is detected.
The on-line visual monitoring system consists of four modules— the image capture module, PLC controlling program, mechanism assembly module, and image inspection software. The image capture module includes a CCD camera、lens、image acquisition board, and lighting to take images. The PLC controlling program receives the signal from the sensor located in the RP machine, commands the image capture module to work, and activates the alarm if mistakes happen. Mechanism assembly module supports and assembles the image capture components in a fixed position. The image inspection software adopts two approaches to inspect captured images—area comparison method and analysis of the interval of point coordinates. Combining these two methods, the system can distinguish the profile geometry and detect the defects effectively. The acceptable proportion value in the area comparison analysis is set between 0.95 and 1.05, and the maximum allowable value in the analysis of the interval of point coordinates is 10 pixels in this research. These values were applied to inspect the designed example model. As a result, the defects were found, and the system triggered the alarm immediately.

[1] Kodama, Hideo , “Automatic method for fabricating a three dimen- sional plastic with photo hardening polymer, ” Review of Scientific Instruments, Volume 52, Number 11, 1981.
[2] Groover P. Mikell, “Fundamentals of Modern Manufacturing ,”John Wiley ＆ Sons, Inc. 2002.
[3] Alejandro Carrion, ”Technology forecast on ink-jet head technology application in rapid prototyping,” Rapid Prototyping Journal Volume 3,Number 3,1997.
[4] Serope Kalpakjian, “ Manufacturing Engineering and Techology ,” Prentice-Hall, Inc. 2001.
[5] Objet Geometries Ltd Company ( http:// www.2objet.com / Products / eden330.html)
[6]林宗德，“新彈性層式及噴霧式製程與線上影像監控之快速原型機研發”，碩士論文，國立臺灣科技大學機械工程系，民國88年6月。
[7]賴後權，“基因演算法於快速成型中網格最佳化與線上影像監視之研究”，碩士論文，大葉大學自動化工程學系，民國92年6月。
[8]汪光夏，“機器視覺應用”，電路版會刊第二期，pp.9~20，1998年10月。
[9] Tokyo Electronic Industry Corporation, “ CCD Monochrome Video Camera CS8420i Operation Manual ”, 2000.
[10] Cox., 攀家寅譯，“攝影光學”，pp.1~10，pp.28~65，徐氏基金會。
[11] National Instruments Corporation, “ IMAQ PCI/PXI – 1411 User Manual ”, 2000.
[12]張笑航，1985，微電腦可程式控制器原理與應用，全華圖書有限公司，pp2~7，pp58~62。
[13]廖文輝，周至宏 編，洪英榮 校閱，“圖形監控”，pp.1-1~1-18， pp.2-1~2-18，全華科技圖書。
[14]洪榮哲，”機械控制用感測器技術入門”，全華科技圖書，pp27~40，1988年10月。
[15] National Instruments Corporation, “IMAQ Vision Concepts Manual”, 2000.
[16]吳成柯，戴善榮，程湘君，雲立貴 譯，“數位影像處理”，pp.5~17，pp.523~533，儒林書局，1996。
[17]劉佳賢, “影像三軸定位系統之研製,” 碩士論文, 國立臺灣科技大學機械工程系, 民國91年6月。