本論文的主要目標為將影像處理與辨識的技術以及雙眼視覺測距應用並實現於互動型教育機器人，使機器人擁有互動與教學的功能。本視覺系統是使用雙攝影機作為輸入裝置，攝影機位置為機器人頭的雙眼，並以一台筆記型個人電腦來作為影像處理中樞。攝影機擷取出的影像畫面，針對不同的目標物使用相對應的影像處理，把目標物的特徵擷取出後，並以此影像特徵來進行影像辨識處理或雙眼視覺測距，再將計算出的目標物三維空間座標，傳遞給機器人，讓機器人可以達成以下互動與教學的功能：1) 模仿使用者人臉表情及上肢姿態 2) 算術教學，讓使用者與機器人能夠進行算術的出題與解答之互動 3) 英文單字教學，由使用者拿取不同的圖卡，使機器人拼出圖卡代表的英文單字。4)頭部追蹤色球並以手碰觸。由實驗結果可知，機器人均能有效成功展示以上提到的功能。 The thesis proposes the techniques of image processing, pattern recognition, and binocular vision distance measure for an interactive educational robot so that teaching and learning can be performed between human and the robot. The robot has two cameras as its two eyes to capture the image and a laptop computer is the center to do the image process and 3D distance measure from the feature of the target so that the robot can achieve teaching, learning, and other interactive motions. Those motions include 1) imitating the user's facial expressions and upper limb posture, 2) teaching users to solve arithmetic questions, 3) recognizing different pictures and spelling the corresponding English words, and 4) tracking and touching a color ball. According to a series of experiences, the robot can complete the above performance successfully.
