Summary: | 碩士 === 銘傳大學 === 電腦與通訊工程學系碩士班 === 102 === The posture recognition system based on Kinect sensor system is developed in this thesis. This system consists of a Kinect posture recognition system, a self-organization map (SOM) algorithm and a flag semaphores learning platform. Although the posture recognition is widely applied for human-robot interaction, there are no uniform postures to be a criterion of control action. The flag semaphores are adopted in this thesis. Then, information of skeleton point is collected by Kinect sensor. The SOM is utilized to analysis the posture information and to recognize the semaphore signal. This thesis will apply these technologies to the development of semaphores teaching and training platform which consists of a letters interactive learning platform, a word interactive learning platform and a semaphores ability identification platform. They will match interactive learning system to guide users to use semaphores posture correctly. In the experiments, we can know that the posture recognition system can be accomplished to recognize the related semaphore signal, and has personal and high scalability advantages.
|