Design and Implementation of Real-Time Stereo-Vision System for Semi-Humanoid Robot

• Main Authors: Bing-Rung Tsai, 蔡秉融
• Other Authors: Tzuu-Hseng Li
• Format: Others
• Language: en_US
• Published: 2008
• Online Access: http://ndltd.ncl.edu.tw/handle/61800713877976384037

碩士 === 國立成功大學 === 電機工程學系碩博士班 === 96 === This thesis proposes a real-time image processing approach for face detection, elevator button recognition, and a stereo vision system for a semi-humanoid robot in an indoor environment. The thesis combines several image processing techniques, which include red region analysis, erosion and dilation, edge-detection, and chain codes to detect the position of the elevator button. The feature of each numeral gap direction is encoded on the button for number recognition and then combined with a three-dimensional coordinate system to acquire its relative coordinate from robot to button. This is done by analyzing the complexion region in the image to find face candidates and then using these characteristics, which are the black and white of the eye and lips as the redder region of the face, for verifying the detected face candidates. Moreover, the position of the person can be obtained in the environment. Finally, the practical experiments demonstrate the feasibility and effectiveness of the proposed schemes.