OpenGL Based Affine SIFT Real-time Image Processing Method for Home Service Robots

• Main Authors: Hao-EnCheng, 鄭浩恩
• Other Authors: Tzuu-Hseng Li
• Format: Others
• Language: en_US
• Published: 2012
• Online Access: http://ndltd.ncl.edu.tw/handle/97233543245028645079

碩士 === 國立成功大學 === 電機工程學系碩博士班 === 100 === This thesis mainly discusses the OpenGL (Open Graphics Library) based affine SIFT (Scale-Invariant Feature Transform) real-time image processing method for home service robots. Firstly, a six-step method is proposed to create image database. When SIFT algorithm is adopted to detect features with this database, the rate of error features is reduced. Integration of the six-step method and SIFT can significantly decrease time for setting up the database. Secondly, the object recognition vision system is built up by parallel Affine-SIFT (ASIFT) algorithm, where OpenGL reduces the calculation time of affine projection and SiftGPU (Graphics Processing Unit) accomplishes the real-time operation speed of SIFT. Thirdly, the human recognition vision system is established by combining face detection of Haar-Like features method, the face recognition of Eigenfaces method and human skeleton detection. Integration of human skeleton and face detection lowers the false rate of facial detection. Adding facial organs detection augments the identification rate of face recognition. Furthermore, a user friendly human-computer interface with OpenGL is constructed and described. Operators can easily obtain the information from the interface such as dimensional image of home environment, values of sensors, and the recognition result of vision system. Finally, both the experimental results in the laboratory and competition consequents of 2012 Robocup@home Japan Open demonstrate the validity, effectiveness, and robustness of the proposed real-time vision system on our home service robot.