本論文在小型人型機器人系統上以擴張型卡爾曼濾波器(Extended Kalman Filter, EKF)實現單眼視覺的同時定位與建圖(Simultaneous Localization And Mapping, SLAM)。應用影像特徵點辨識與追蹤方法,找到影像特徵點,利用2張影像中相同的特徵點計算深度,得知特徵點在空間中的3D座標,提供機器人EKF SLAM之用。本研究也自製一部符合RoboCup小型(kid-szie)規格的人型機器人,搭配單板電腦(PC-based)控制器。整合機器人與EKF SLAM系統,將能夠在簡易的室內環境中,進行同時自我定位與地圖建立。 In this thesis, a single-camera simultaneous localization and mapping (SLAM) algorithm based on the extended kalman filter(EKF) is developed for a RoboCup Kidsize humanoid robot. In order to find the image-features, we applied the recognition of features and the tracking techniques, using two same image-features to calculate the depth. The three dimension feature positions would be used by robot EKF SLAM.We design and fabricate a PC-based humanoid robot which is conformed to the regulation of RoboCup Kidsize humanoid robot league. The integrated humanoid robot and EKF SLAM system can be implemented SLAM in simple indoor environments.