本研究將以影像區域特徵(image region features)追蹤與偵測方法規劃機器人的感測模型，提供人形機器人系統同時建立環境地圖與自我定位、以及物件方向與位置偵測之用。本研究的內容分為三個部分：包括影像區域特徵的偵測與追蹤，特徵初始化(feature initialization)程序的規劃，機器人同時定位與建圖、以及移動物件方位偵測的應用等。發展的視覺模組應用在具備前視型攝影機的小型人形機器人系統上。 In this thesis, the method of image feature detecting and tracking is utilized to develop the sensor module of a robot, and provided for robot simultaneous self-localization and mapping, as well as detection of the position and orientation of an object. The research is divided into three stages: including the detection and tracking of image region features; the algorithm of feature initialization; and the task of simultaneous localization and mapping, as well as the detection of the position and orientation for a moving object. The developed vision module is applied to a small-size humanoid robot system with a frontal camera.
