本論文主要提供影像處理及辨識之技術並實現於一部輪型機器人，使之具有自主跟隨與自動搭乘電梯的功能。在自主跟隨部分，藉由各種感測器資訊 (紅外線、超音波、光球裝置) 及無線通訊模組 (Zigbee)，搭配影像處理技術與控制策略，讓機器人能夠即時辨識及跟隨主人，並保持適當距離。機器人利用光球裝置能夠變色的功能，搭配Zigbee即時傳送訊息，即使存在有其他 (相同或相異) 光源干擾，依然能夠成功辨識出主人。在辨識電梯內外環境與自動搭乘部分，主要利用攝影機擷取環境影像 (包括電梯門、電梯面板、電梯按鈕)，再透過影像辨識技術，並結合感測器資訊 (紅外線與超音波)，讓機器人能夠達到以下功能：1) 正確判斷外部電梯門位置； 2) 自動偵測內部電梯面板； 3) 正確辨識內部電梯按鈕，並成功按壓。自主跟隨與自動搭乘電梯的實驗最後將成功呈現於本論文之第五章。The thesis proposes an image processing technique on a wheel mobile robot to achieve the functions of people following and elevator taking. Concerning the people following, by using the technology of image processing and control strategy, the robot can recognize the master and keep in an appropriate distance between master and robot and follow the master in real time according to the information of sensors adopted from infrared, sonar, wireless communication module (Zigbee) and the device of light. In addition, by using Zigbee, the robot can communicate with the light ball, which is tie in the back of the master, to change its color when there is other same color lights appearing in the vision of the robot such that the master can be recognized continuously. Concerning the elevator taking, camera captures the image of the elevator environment, including elevator door, elevator panel and elevator button. Using image processing technique and sensors (infrared and sonar) information, the robot can achieve the tasks as follows. 1) Detecting the position of elevator door outside; 2) detecting the position of elevator panel inside; 3) recognizing and pushing the elevator button correctly. Therefore, the robot can recognize, follow people and take elevator successfully. The experiment result is implemented successfully and shown in Chapter 5 of this thesis.
