本文主要目的是利用影像辨識樂譜上之音符音階，並以特殊之機構設計演奏 實體樂器，完成自動演奏機器人的開發。文中以圖控軟體LabVIEW 完成光學字元 辨識演算與影像辨識，並將辨識結果轉換為相對應的馬達訊號，透過Arduino 控制 伺服馬達，完成樂器的吹奏。同時特別設計馬達帶動仿製手指模擬人類吹笛的方 式吹奏直笛，完成即時辨識樂譜並立即吹奏的功能。實驗結果證實影像辨識技術 可以實現自動吹笛音樂合之功能。將來之應用可推廣至教具開發與其他智慧型自 動樂器之開發。

This article is aimed to develop an autonomous performing robot using image recognition technology to identify musical notes and musical scales on music scores such that a real musical instrument can be played through a special-designed structure. In order to play the recorder, LabVIEW system design software will be used to conduct optical character recognition (OCR) calculation and image recognition. Meanwhile, the recognition results will be translated into corresponding signals such that the servo motor can be controlled by Arduino microchips. Moreover, customized humanoid fingers are designed to imitate the human way to play a recorder. The experimental results indicate the practicality of the image recognition for autonomous fluted music boxes. In the future, this work can be extended to the development of teaching aids and other intelligent automatic musical instruments.