本研究整合機器視覺 (Machine vision) 與微製造技術 (Micro fabricationtechnique) 功能於同一部製造系統上，利用精密位移平台將線上加工完成的微細元件，以影像疊合技術對誤差量進行線上補償檢測，透過影像擷取、濾波、二值化、影像定位及影像比對的方式，計算工件影像和標準影像像素座標差值，即可得知補償值，經影像辨識計算後的補償量，系統會自動補償於原始加工程式的原點座標 (G92) 上進行再加工。主要目的使微小加工元件，在沒有任何移動、拆卸的情況下，直接進行影像比對與補償再加工，以提高微元件之幾何精度與尺寸精度，節省校正時間，並獲致線上微細加工、檢測與自動化補償之目的。透過上述所開發的補償技術，本研究成功應用於直徑30 μm 精密微球探針的製作。

This study combines “machine vision” with “micro fabrication technique” in a developed tabletop machine tool to construct a technique of on-line compensation and re-machining via sensing image information. A micro globular probe with 30 μm in diameter is designed and rough machined using micro electrical discharge machining from a micro rod made of tungsten carbide. Subsequently, it is cleaned on-line using the ultrasonic vibration-assisted in the designed cleaning tank. Image recognition which incorporates image capturing, binary, edge detecting and noise filtering as well as image superimposing is real-time carried out after rough machining to determine the values of error compensation and offset the coordinates of G92 command during finish machining. Experimental results show a high dimensional accuracy, geometry accuracy
and excellent surface roughness can be obtained. The proposed approach is very cost-effective and can significantly contribute to the high precision on-line micro measurement industry.