Development of a Vision System for Humanoid Robot

• Main Authors: Qusuw Pusin, 奎樹‧布興
• Other Authors: Chi-Sheng Tsai
• Format: Others
• Language: zh-TW
• Online Access: http://ndltd.ncl.edu.tw/handle/2b2v9e

碩士 === 建國科技大學 === 自動化工程系暨機電光系統研究所 === 96 === It is necessary for the humanoid robot with perceptive systems to make judgments in motions and actions. The vision system is one of efficient perceptive systems. Thus, the objective of this thesis is to develop a vision system for the humanoid robot to capture the related information that occurred in its surrounding environment. The function of this developed vision system consists of line, face and object detections. In line detection module, after the use of smoothing to improve the effect of line and noise, the Canny Edge detection has been applied to obtain the candidate points of line. Finally, the Hough Transformation is introduced to find the line. In face detection module, after the use of skin color detection, the color image is converted to gray image. Then, Binary operator and Morphology are introduced to deal with the portrait. These processes are taken to be the bases for identifying the eye and mouth of the face. The color segmentation, text finding, and the corner detection are used in the object detection to measure the distance of the object. A prototype of the system has been built and tested. The results reveal that the developed system is able to detect line, face and object, and to feed back the detected information to related system in the humanoid robot.